CN105184250A - Ground object classification method based on electric power corridor airborne LiDAR point cloud data - Google Patents
Ground object classification method based on electric power corridor airborne LiDAR point cloud data Download PDFInfo
- Publication number
- CN105184250A CN105184250A CN201510547324.1A CN201510547324A CN105184250A CN 105184250 A CN105184250 A CN 105184250A CN 201510547324 A CN201510547324 A CN 201510547324A CN 105184250 A CN105184250 A CN 105184250A
- Authority
- CN
- China
- Prior art keywords
- point
- point cloud
- cloud data
- feature
- electric power
- 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.)
- Granted
Links
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
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
- G06F18/24143—Distances to neighbourhood prototypes, e.g. restricted Coulomb energy networks [RCEN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/254—Fusion techniques of classification results, e.g. of results related to same input data
- G06F18/256—Fusion techniques of classification results, e.g. of results related to same input data of results relating to different input data, e.g. multimodal recognition
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Multimedia (AREA)
- Traffic Control Systems (AREA)
- Image Processing (AREA)
- Optical Radar Systems And Details Thereof (AREA)
Abstract
The invention provides a ground object classification method based on electric power corridor airborne LiDAR point cloud dada, and relates to the airborne LiDAR point cloud data processing field. In order to solve the problem of insufficient utilization of information included in the electric power corridor airborne LiDAR point cloud dada in the prior art, the invention provides a ground object classification method based on electric power corridor airborne LiDAR point cloud data features. The method comprises the following steps: obtaining airborne LiDAR point cloud dada; removing gross error points of the point cloud dada; carrying out point cloud dada feature extraction and processing, wherein the point cloud dada features comprise single point feature and neighbourhood feature, and carrying out normalization processing on the extracted features; and selecting samples having labels from a LiDAR point cloud dada set, separating target information of power lines, vegetation, buildings and earth surface and the like in the LiDAR point cloud dada, and classifying the data set by utilizing the samples having labels to obtain class information of power lines, vegetation, buildings and earth surface and the like.
Description
Technical field
The present invention relates to airborne LiDAR point cloud data processing field; In particular to the terrain classification method of a kind of electric power corridor airborne LiDAR point cloud data.
Background technology
The general visual interpretation relying on human eye is patrolled and examined in traditional electric power corridor, and except at substantial manpower and materials, the data precision obtained is not high yet, and for complicated landform, vehicle, manpower are difficult to arrive, and patrol and examine bring larger difficulty to electric power corridor.And adopting helicopter to carry the equipment such as digital camera, infrared video camera, the data deficiency three-dimensional information obtained, can not judge the atural object status information in electric power corridor exactly.And airborne LiDAR (LightDetectionandRanging) system can the three-dimensional spatial information of quick obtaining ground scene, has high-level efficiency, low-loss, high-precision feature, in therefore starting in recent years more to patrol and examine for electric power corridor.
Utilize airborne LiDAR to patrol and examine electric power corridor at present, the large multipair electric power facility of the LiDAR point cloud data obtained (comprising electric wire and electric tower) carries out three-dimensional modeling.And not only comprise electric power facility information in the cloud data that airborne LiDAR system obtains, also comprise other terrestrial object informations in electric power corridor, comprise vegetation, building etc.If only pay close attention to electric power facility point cloud, the utilization ratio of LiDAR point cloud data is reduced greatly.A kind of method based on a cloud level journey threshold filter and density filtering is disclosed in the application for a patent for invention " a kind of airborne LIDAR data line of electric force rapid extraction and reconstructing method " of publication number CN102590823A disclosed in 18 days July in 2012, line of electric force and electric tower point cloud is extracted from LiDAR point cloud, then carry out layering matching to a cloud and obtain single power line vector, the line of electric force achieved between multiple span automatically extracts and extracts with electric tower.Publication number disclosed in 16 days January in 2013 is disclose a kind of power line extraction method based on carrying out matching to the electric power cluster center of xsect sampled point cloud cluster in the application for a patent for invention " a kind of power line extraction method of the transmission line of electricity based on on-board LiDAR data " of CN102879788A, copies and skew obtains multiple-loop line road electric wire to the center line after matching.Publication number disclosed in 27 days November in 2013 is disclose a kind of method utilizing LiDAR point cloud to carry out electric tower extraction in the application for a patent for invention " automatically extracting method of power tower in random laser point cloud data " of CN103412296A.These technology are all the extraction and the reconstruction that utilize airborne LiDAR point cloud data to carry out line of electric force or electric tower in electric power corridor, ignore a large amount of useful informations comprised in non-electricity facility point cloud.
Summary of the invention
The present invention is only limitted to the extraction to line of electric force in order to solve prior art to airborne LiDAR point cloud data in electric power corridor, makes the problem of comprised Information Pull deficiency, therefore proposes the terrain classification method of a kind of electric power corridor airborne LiDAR point cloud data.
Airborne LiDAR system is made up of GPS positioning system, INS inertial navigation system and laser scanning and ranging system three subsystems.Wherein, GPS positioning system realizes the location to airborne platform; The attitude information of INS inertial navigation system measuring table, comprises roll angle, the angle of pitch and position angle; The scanning angle of laser scanning and ranging system log (SYSLOG) laser and flight time, obtain the distance of culture point apart from laser receiver by measuring the flight time.Concrete steps are as follows:
LiDAR point cloud data are carried out three-dimensional visualization display by step one, acquisition LiDAR point cloud data, and setting elevation threshold value removes rough error point;
Step 2, feature extraction and characteristic processing are carried out to LiDAR point cloud data;
Step 3, from LiDAR point cloud data, Stochastic choice has exemplar;
Step 4, utilization have exemplar to be classified to LiDAR point cloud data sample to be sorted by sorter and k neighbour criterion sorting technique, obtain electric power facility, vegetation, buildings, earth's surface cloud data;
Step 1, calculate sample to be sorted and all theorem in Euclid space distances having exemplar
y represents sample to be sorted, x
iindicate exemplar, n
irepresent that the i-th class has the quantity of exemplar;
Step 2, according to k neighbour criterion, sample to be sorted to be classified, by sample to be sorted with have exemplar to be divided into a class apart from minimum, namely k arest neighbors of sample to be sorted has the sample size of the i-th class in exemplar to be m
i, then class label ω=argmaxm
i.
Wherein step one detailed process is:
Step one by one, utilize MATLAB software that cloud data is read workspace, and carry out visual display;
Step one two, basis are not less than ground, not higher than ground object target the highest in scene point, determine the elevation interval [L that atural object impact point exists, H], wherein L represents elevation threshold value minimum point, and H represents elevation threshold value peak, elevation threshold value is less than the point that L and elevation threshold value be greater than H and weeds out.
Step 2 detailed process is:
Step 2 one, extraction single-point feature;
Step 2 two, extraction neighborhood characteristics;
Step 2 three, feature normalization and feature selecting.
The height of step 2 one by one, using in LiDAR point cloud, intensity, echo times, echo position information are all as a feature of current point;
Step 2 one or two, LiDAR point cloud is carried out piecemeal, the minimum point chosen in every block carries out terrain model matching as ground point, original point cloud level degree is deducted ground level and obtains relative altitude information a little as a feature; Point relative height being less than 1m is set as ground point.
Distance between two points in step 221, calculation level cloud, finds out and current point P
0k nearest point forms its neighbour's point set { P
i, i=1,2 ..., k, the 3 d space coordinate vector representation of often is P
i=(x
i, y
i, z
i)
t; Wherein, P
irepresent i-th Neighbor Points vector, x
i, y
i, z
irepresentation space three-dimensional coordinate, T represents transposition.
Step 2 two or two, calculate the covariance matrix that each point concentrates all somes three-dimensional coordinates
wherein,
represent k Neighbor Points mean value, P
0represent current point; Solve three eigenvalue λ of Σ
1> λ
2> λ
3and three of correspondence proper vector v
1, v
2, v
3.
Using three features of eigenwert as current point, eigenwert basis defines other characteristic quantities:
Subjunctive vector is v
3=(v
31, v
32, v
33)
t, then angle is
using the feature of angle angle information as current point.
Height variance, intensity variance, the angle variance of point in step 2 two or three, calculating neighborhood, as three features of current point;
Step 2 two or four, search neighborhood point set middle distance point farthest, according to the dot density feature of the spheroid volume computing neighborhood point set using this maximum distance as radius.
Step 231, the feature cloud data extracted, utilize formula f=(f-min)/(max-min), f ∈ [0,1], and wherein max is the maximal value of feature f, and min is the minimum value of feature f;
Step 2 three or two, character numerical value after normalization concentrate near 0 or 1, and dynamic range is less; And residue character numeric distribution after normalization is more discrete, dynamic range is comparatively large, then by y=ln (x+1) or y=e
x+1dynamic range is adjusted.
Step 3 detailed process is:
Step 3 one, by determining certain point or classification information of certain point set, being added to corresponding has in exemplar set;
Step 3 two, selected by have exemplar to be uniformly distributed and sorted Different categories of samples number is equal.
Beneficial effect of the present invention:
The invention solves prior art electric power corridor scene airborne LiDAR point cloud data processing technique and can not obtain Appropriate application, with electric power corridor area airborne LiDAR point cloud data for research object, propose a kind of terrain classification method of new electric power corridor airborne LiDAR point cloud data.By the acquisition of airborne LiDAR point cloud data; Cloud data rough error point is removed; Cloud data feature extraction and process, its point cloud data feature comprises single-point characteristic sum neighborhood characteristics, and is normalized the feature extracted; From LiDAR point cloud data acquisition, select there is exemplar, there is exemplar to classify to data set target information utilizations such as isolating line of electric force, vegetation, building and earth's surface in LiDAR point cloud, obtain the classification information such as line of electric force, vegetation, buildings and earth's surface.In LiDAR point cloud data, information obtains fully effectively utilizing.The present invention is applied to field of three-dimension modeling.
Accompanying drawing explanation
Fig. 1 is the FB(flow block) schematic diagram of the terrain classification of electric power corridor of the present invention scene airborne LiDAR point cloud data;
Fig. 2 is that in embodiment, airborne LiDAR system obtains electric power corridor scene point cloud data plot, and Fig. 2 (a) is initial three-dimensional point cloud chart, and Fig. 2 (b) is three-dimensional point cloud side view, and Fig. 2 (c) is embodiment cloud data figure used;
Fig. 3 is rough error point schematic diagram in LiDAR point cloud in embodiment: Fig. 3 (a) is 3-D view, and Fig. 3 (b) is side view, and wherein, zero represents flying spot, and represents low spot;
Fig. 4 is that in embodiment, LiDAR point cloud rough error point rejects design sketch: Fig. 4 (a) is original point cloud three-dimensional effect diagram, Fig. 4 (b) is that rough error point rejects result three-dimensional effect diagram, Fig. 4 (c) is original point cloud section design sketch, Fig. 4 (d) is that rough error point rejects result section design sketch, Fig. 4 (e) is original point cloud vertical view, and Fig. 4 (f) rough error point rejects result vertical view;
Fig. 5 is single-point feature information extraction result in embodiment: Fig. 5 (a) is altitude feature, and Fig. 5 (b) is strength characteristic, and Fig. 5 (c) is echo times feature;
Fig. 6 is embodiment mesorelief modeling schematic diagram: Fig. 6 (a) is original point cloud, and Fig. 6 (b) is that ground Seed Points chooses result, and Fig. 6 (c) is terrain model modeling result;
Fig. 7 is that in embodiment, neighborhood choice schematic diagram: Fig. 7 (a) represents spheric neighbo(u)rhood N
3Dwith cylindricality neighborhood N
2D, Fig. 7 (b) represents k neighbour point set neighborhood;
Fig. 8 is LiDAR point cloud geometry feature extraction schematic diagram in embodiment: Fig. 8 (a) represents original point cloud, and Fig. 8 (b) represents planar structure feature extraction result, and Fig. 8 (c) represents that linear structural feature extracts result;
Fig. 9 is that in embodiment, characteristic dynamic range Tuning function schematic diagram: Fig. 9 (a) is applicable to the situation that character numerical value concentrates on smaller value, and Fig. 9 (b) is applicable to the situation that character numerical value concentrates on higher value;
Figure 10 is that in embodiment, before and after characteristic dynamic range adjustment, comparison diagram: Figure 10 (a) represents certain primitive character, Figure 10 (b) represents the feature after dynamic range adjustment, Figure 10 (c) represents the plan view of primitive character, and Figure 10 (d) represents the rear characteristic plane view of dynamic range adjustment;
Figure 11 is that embodiment point cloud classifying quality figure: Figure 11 (a) represents original point cloud, Figure 11 (b) represents line of electric force, Figure 11 (c) represents vegetation, and Figure 11 (d) represents buildings, and Figure 11 (e) represents Ground Point.
Embodiment
Embodiment one: composition graphs 1 illustrates present embodiment, the terrain classification method of the electric power corridor on-board LiDAR data of present embodiment comprises the following steps:
LiDAR point cloud data are carried out three-dimensional visualization display by step one, acquisition LiDAR point cloud data, remove rough error point according to setting elevation threshold value;
Step 2, feature extraction and characteristic processing are carried out to LiDAR point cloud data;
Step 3, from LiDAR point cloud data, Stochastic choice has exemplar;
Step 4, utilization have exemplar to be classified to LiDAR point cloud data sample to be sorted by sorter and k neighbour criterion sorting technique, obtain electric power facility, vegetation, buildings, earth's surface cloud data;
Step 1, calculate sample to be sorted and all theorem in Euclid space distances having exemplar
y represents sample to be sorted, x
iindicate exemplar, n
irepresent that the i-th class has the quantity of exemplar;
Step 2, according to k neighbour criterion, sample to be sorted to be classified, by sample to be sorted with have exemplar to be divided into a class apart from minimum, namely k arest neighbors of sample to be sorted has the sample size of the i-th class in exemplar to be m
i, then class label ω=argmaxm
i.
The beneficial effect of present embodiment:
The invention solves prior art electric power corridor scene airborne LiDAR point cloud data processing technique and can not obtain Appropriate application, with electric power corridor area airborne LiDAR point cloud data for research object, propose a kind of terrain classification method of new electric power corridor airborne LiDAR point cloud data.
Embodiment two: the difference of present embodiment and embodiment one is: the removal rough error point of present embodiment described in step one, concrete steps are as follows:
Step one by one, utilize MATLAB software that cloud data is read workspace, and carry out visual display;
Step one two, basis are not less than ground, not higher than ground object target the highest in scene point, determine the elevation interval [L that atural object impact point exists, H], wherein L represents elevation threshold value minimum point, and H represents elevation threshold value peak, elevation threshold value is less than the point that L and elevation threshold value be greater than H and weeds out.
Embodiment two: the difference of present embodiment and embodiment one is: present embodiment carries out feature extraction and characteristic processing described in step 2, and concrete steps are as follows:
Step 2 one, extraction single-point feature;
Step 2 two, extraction neighborhood characteristics;
Step 2 three, feature normalization and feature selecting.
Embodiment four: the difference of present embodiment and embodiment three is: the extraction single-point feature of present embodiment described in step 2 one, concrete steps are as follows:
The height of step 2 one by one, using in LiDAR point cloud data, intensity, echo times, echo position information are as a feature of point;
Step 2 one or two, LiDAR point cloud data are carried out piecemeal, the minimum point chosen in every block carries out terrain model matching as ground point, original point cloud level degree is deducted ground level and obtains relative altitude information a little as a feature; What relative height is less than 1m is set as ground point.
Embodiment five: the difference of present embodiment and embodiment three is: the extraction neighborhood characteristics of present embodiment described in step 2 two, concrete steps are as follows:
Distance between two points in step 221, calculation level cloud, finds out and current point P
0k nearest point forms its neighbour's point set { P
i, i=1,2 ..., k, the 3 d space coordinate vector representation of often is P
i=(x
i, y
i, z
i)
t, wherein, P
irepresent i-th Neighbor Points vector, x
i, y
i, z
irepresentation space three-dimensional coordinate, T represents transposition;
Step 2 two or two, calculate the covariance matrix that each point concentrates all somes three-dimensional coordinates
wherein,
represent k Neighbor Points, P
0represent current point; Solve three eigenvalue λ of Σ
1> λ
2> λ
3and three of correspondence proper vector v
1, v
2, v
3;
Using three features of eigenwert as current point, eigenwert basis defines other characteristic quantities:
According to the physical significance of eigenwert, proper vector, the 3rd proper vector v
3the normal vector of corresponding fit Plane, and on the basis of normal vector computing method vector v
3with the angle of vertical direction, using angle angle information as current point P
0a feature; Subjunctive vector is v
3=(v
31, v
32, v
33)
t, then angle is
Height variance, intensity variance, the angle variance of point in step 2 two or three, calculating neighborhood, as three features of current point;
Step 2 two or four, search neighborhood point set middle distance point farthest, according to the dot density feature of the spheroid volume computing neighborhood point set using this maximum distance as radius.
Embodiment six: the difference of present embodiment and embodiment three is: present embodiment is when the feature normalization described in step 2 three and feature selecting, and concrete steps are as follows:
Step 231, the feature cloud data extracted, utilize formula f=(f-min)/(max-min), f ∈ [0,1], and wherein max is the maximal value of feature f, and min is the minimum value of feature f;
Step 2 three or two, character numerical value after normalization concentrate near 0 or 1, and dynamic range is less; And residue character numeric distribution after normalization is more discrete, dynamic range is comparatively large, then by y=ln (x+1) or y=e
x+1dynamic range is adjusted.
Embodiment seven: the difference of present embodiment and embodiment one is: present embodiment is when the selection described in step 3 has exemplar, and concrete steps are as follows:
Step 3 one, by determining certain point or classification information of certain point set, being added to corresponding has in exemplar set;
Step 3 two, selected by have exemplar to be uniformly distributed and sorted Different categories of samples number is equal.
Embodiment
Composition graphs 1-11 illustrates the present embodiment, the present embodiment is the terrain classification method of a kind of electric power corridor airborne LiDAR point cloud data characteristics, electric power corridor airborne LiDAR point cloud data shown in main employing Fig. 2 are as a raw data for cloud feature extraction and classifying, this raw data is that the cloud data in the Sonoma area that intercepting provides from Univ Maryland-Coll Park USA is (from " SonomaCountyVegetationMappingandLiDARProgram " project, CarbonMonitoringSystem by NASA subsidizes), as shown in Fig. 2 (c), point cloud sum about 5,800,000, point cloud density is 13.75 points/m
2, used cloud quantity about 100,000 point of intercepting.
The present embodiment airborne LiDAR point cloud data characteristics is extracted and is all carried out on Matlab platform with the realization of classification, and as shown in Figure 1, concrete implementation step is as follows for concrete implementing procedure:
1. digital independent---the LiDAR point cloud data downloaded are txt form, and source document comparatively large (about 470M), namely MATLAB software is adopted to read in workspace by cloud data, be converted to mat data layout, wherein the information such as X-coordinate, Y-coordinate, height value, echo strength, echo times of each column data representation space point successively.
2. a cloud elimination of rough difference---after reading in cloud data, it is shown with three dimensions form, as Fig. 2 (a), according to its elevation distribution situation, as shown in Fig. 2 (b), determine that elevation interval is [135m, 308m] artificially, by height higher than 308m or highly lower than the rough error point deletion of 135m; Embodiment has intercepted one piece of region and has processed, as shown in Fig. 2 (c) in the middle of the cloud data after excluding gross error point; Shown in Fig. 3, rough error point schematic diagram; As shown in Figure 4, the contrast of characteristic before and after rough error point is rejected: Fig. 4 (a) and Fig. 4 (b) are that original point cloud three-dimensional effect diagram and rough error point are rejected result three-dimensional effect diagram and contrasted; Fig. 4 (c) and Fig. 4 (d) is that original point cloud section design sketch and rough error point reject result section design sketch; Fig. 4 (e) and Fig. 4 (f) is that original point cloud vertical view and rough error point reject result vertical view.
3. in single-point feature extraction---the cloud data that airborne LiDAR system obtains, not only comprise the three dimensional space coordinate information of often, also comprise the information such as intensity, number of times, position about return laser beam, the elevation information of a cloud, strength information, echo times information are stored as single-point characteristic information, these three single-point characteristic informations are respectively as shown in Fig. 5 (a), 5 (b) He 5 (c).
Elevation information represents the height of the three-dimensional point that data that GPS, INS in LiDAR system, laser scanning system obtain calculate through coordinate, the return laser beam intensity that strength information is obtained by LiDAR system obtains through quantification, and the return laser beam crest quantity that echo times is received by LiDAR receiving system is determined.
For relative height feature, need to set up ground level model, and original point cloud is as contained buildings in Fig. 6 (a) scene, therefore when choosing ground Seed Points, setting sizing grid is 30m, can comprise ground point in each mesh scale, finally obtains 144 ground Seed Points altogether, be evenly distributed in the plane of about 360m × 360m, ground Seed Points is as Fig. 6 (b).Utilize surface fitting function griddata to carry out modeling to ground level to all Seed Points, terrain model is as shown in Fig. 6 (c).Deduct ground level model with pilot region original point cloud level degree and obtain relative height feature.
4. neighborhood characteristics is extracted---and some cloud variable density makes the some quantity in fixed measure neighborhood differ greatly, the lazy weight even put is to carry out local space structure analysis, therefore the present embodiment adopts k neighbour point set to substitute fixed measure neighborhood to carry out neighborhood characteristics extraction, selected Neighbor Points quantity is 10, as shown in Figure 7.Calculate the three-dimensional coordinate covariance matrix of point set
(wherein,
p
0represent current point) and eigenwert, proper vector, planar process vector angle.Eigenwert is utilized to calculate partial structurtes information, as shown in Fig. 8 (b), (c).Calculate echo strength Variance feature
with height variance feature
wherein, I
0, h
0represent echo strength information and the elevation information of current point, h
i,
represent in neighborhood point set often height and average height a little, I
i,
represent in neighborhood point set often echo strength and mean echo intensity a little.
5. every one-dimensional characteristic f that feature pre-service---step 4 is extracted, all utilize formula f=(f-min)/(max-min) to be transformed in [0,1] interval, wherein max is the maximal value of feature f, and min is the minimum value of feature f.Most of characteristic value set after normalization is near 0 or 1, for maximal eigenvector, its distribution is as shown in Figure 10 (a), major part characteristic point concentrates near 0 or 1, dynamic range is less, and remaining characteristic numeric distribution is more discrete, dynamic range is larger, the function y=ln (x+1) then obtained to left unit by function Fig. 9 (a) Suo Shi is adjusted dynamic range, expand the dynamic range near numerical value 0, dynamic range near compression 1 value, the result obtained is as shown in Figure 10 (b), Figure 10 (c) and 10 (d) are its plane comparison diagrams.
6. point cloud classifications---selecting 50 in the data has exemplar as a class, calculate the theorem in Euclid space distance of sample to be sorted and the 4 classes proper vector of totally 200 exemplar, find out 20 nearest points, the quantity belonging to which kind of point in these 20 points is many, so just sample to be sorted is assigned in the middle of such, classification results as shown in figure 11.
Claims (7)
1. a terrain classification method for electric power corridor airborne LiDAR point cloud data, is characterized in that: a kind ofly comprise the steps: based on the terrain classification method of airborne LiDAR system for electric power corridor LiDAR point cloud data
LiDAR point cloud data are carried out three-dimensional visualization display by step one, acquisition LiDAR point cloud data, and setting elevation threshold value removes rough error point;
Step 2, feature extraction and characteristic processing are carried out to LiDAR point cloud data;
Step 3, from LiDAR point cloud data, Stochastic choice has exemplar;
Step 4, utilization have exemplar to be classified to LiDAR point cloud data sample to be sorted by sorter and k neighbour criterion sorting technique, obtain electric power facility, vegetation, buildings, earth's surface cloud data;
Step 1, calculate sample to be sorted and all theorem in Euclid space distances having exemplar
y represents sample to be sorted, x
iindicate exemplar, n
irepresent that the i-th class has the quantity of exemplar;
Step 2, according to k neighbour criterion, sample to be sorted to be classified, by sample to be sorted with have exemplar to be divided into a class apart from minimum, namely k arest neighbors of sample to be sorted has the sample size of the i-th class in exemplar to be m
i, then class label ω=argmaxm
i.
2. the terrain classification method of a kind of electric power corridor according to claim 1 airborne LiDAR point cloud data, is characterized in that: the concrete steps of described step one are as follows:
Step one by one, utilize MATLAB software that cloud data is read workspace, carry out visual display;
Step one two, basis are not less than ground point, not higher than the ground object target point of peak in scene, determine the elevation interval [L that atural object impact point exists, H], wherein L represents elevation threshold value minimum point, H represents elevation threshold value peak, elevation threshold value is less than the point that L and elevation threshold value be greater than H and weeds out, complete and remove rough error point.
3. the terrain classification method of a kind of electric power corridor according to claim 1 airborne LiDAR point cloud data, is characterized in that: the concrete steps that described step 2 carries out feature extraction and characteristic processing to LiDAR point cloud data are as follows:
Step 2 one, extraction single-point feature;
Step 2 two, extraction neighborhood characteristics;
Step 2 three, feature normalization and feature selecting.
4. the terrain classification method of a kind of electric power corridor according to claim 3 airborne LiDAR point cloud data, is characterized in that: the concrete steps that described step 2 one extracts single-point feature are as follows:
The height of step 2 one by one, using in LiDAR point cloud data, intensity, echo times, echo position information are as a feature of point;
Step 2 one or two, LiDAR point cloud data are carried out piecemeal, the minimum point chosen in every block carries out terrain model matching as ground point, original point cloud level degree is deducted ground level and obtains relative altitude information a little as a feature; What relative height is less than 1m is set as ground point.
5. the terrain classification method of a kind of electric power corridor according to claim 3 airborne LiDAR point cloud data, is characterized in that: the concrete steps that described step 2 two extracts neighborhood characteristics are as follows:
Distance between two points in step 221, calculation level cloud, finds out and current point P
0k nearest point forms its neighbour's point set { P
i, i=1,2 ..., k, the 3 d space coordinate vector representation of often is P
i=(x
i, y
i, z
i)
t, wherein, P
irepresent i-th Neighbor Points vector, x
i, y
i, z
irepresentation space three-dimensional coordinate, T represents transposition;
Step 2 two or two, calculate the covariance matrix that each point concentrates all somes three-dimensional coordinates
wherein,
represent k Neighbor Points mean value, P
0represent current point; Solve three eigenvalue λ of Σ
1> λ
2> λ
3and three of correspondence proper vector v
1, v
2, v
3;
Using three features of eigenwert as current point, eigenwert basis defines other characteristic quantities:
Linear linearity
unchangeability omnivariance
Face shape planarity
anisotropy anisotropy
Spherical sphericity
intrinsic entropy eigenentropy
According to the physical significance of eigenwert, proper vector, the 3rd proper vector v
3the normal vector of corresponding fit Plane, and on the basis of normal vector computing method vector v
3with the angle of vertical direction, using angle angle information as current point P
0a feature; Subjunctive vector is v
3=(v
31, v
32, v
33)
t, then angle is
Height variance, intensity variance, the angle variance of point in step 2 two or three, calculating neighborhood, as three features of current point;
Step 2 two or four, search neighborhood point set middle distance point farthest, according to the dot density feature of the spheroid volume computing neighborhood point set using this maximum distance as radius.
6. the terrain classification method of a kind of electric power corridor according to claim 3 airborne LiDAR point cloud data, is characterized in that: the concrete steps of described step 2 three feature normalization and feature selecting are as follows:
Step 231, the feature cloud data extracted, utilize formula f=(f-min)/(max-min), f ∈ [0,1], and wherein max is the maximal value of feature f, and min is the minimum value of feature f;
Step 2 three or two, character numerical value after normalization concentrate near 0 or 1, and dynamic range is less; And residue character numeric distribution after normalization is more discrete, dynamic range is comparatively large, then by y=ln (x+1) or y=e
x+1dynamic range is adjusted.
7. the terrain classification method of a kind of electric power corridor according to claim 1 airborne LiDAR point cloud data, is characterized in that: in described step 3 LiDAR point cloud data, Stochastic choice has the concrete steps of exemplar as follows:
Step 3 one, by determining certain point or classification information of certain point set, being added to corresponding has in exemplar set;
Step 3 two, selected by have exemplar to be uniformly distributed and sorted Different categories of samples number is equal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510547324.1A CN105184250B (en) | 2015-08-31 | 2015-08-31 | A kind of terrain classification method of electric power corridor airborne LiDAR point cloud data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510547324.1A CN105184250B (en) | 2015-08-31 | 2015-08-31 | A kind of terrain classification method of electric power corridor airborne LiDAR point cloud data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105184250A true CN105184250A (en) | 2015-12-23 |
| CN105184250B CN105184250B (en) | 2018-06-22 |
Family
ID=54906319
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510547324.1A Active CN105184250B (en) | 2015-08-31 | 2015-08-31 | A kind of terrain classification method of electric power corridor airborne LiDAR point cloud data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105184250B (en) |
Cited By (26)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105701478A (en) * | 2016-02-24 | 2016-06-22 | 腾讯科技（深圳）有限公司 | Method and device for extraction of rod-shaped ground object |
| CN105701856A (en) * | 2016-01-15 | 2016-06-22 | 厦门大学 | Vegetation extraction method and system |
| CN106199557A (en) * | 2016-06-24 | 2016-12-07 | 南京林业大学 | A kind of airborne laser radar data vegetation extracting method |
| CN106526610A (en) * | 2016-11-04 | 2017-03-22 | 广东电网有限责任公司电力科学研究院 | Power tower automatic positioning method and apparatus based on unmanned aerial vehicle laser point cloud |
| CN106897686A (en) * | 2017-02-19 | 2017-06-27 | 北京林业大学 | A kind of airborne LIDAR electric inspection process point cloud classifications method |
| CN107273902A (en) * | 2017-05-19 | 2017-10-20 | 中国科学院遥感与数字地球研究所 | A kind of method that electric tower point cloud is automatically extracted from on-board LiDAR data |
| CN107767389A (en) * | 2017-10-27 | 2018-03-06 | 云南电网有限责任公司电力科学研究院 | A kind of power transmission line extraction method based on airborne laser point cloud data |
| CN108562885A (en) * | 2018-03-29 | 2018-09-21 | 同济大学 | A kind of ultra-high-tension power transmission line airborne LiDAR point cloud extracting method |
| CN108564650A (en) * | 2018-01-08 | 2018-09-21 | 南京林业大学 | Shade tree target recognition methods based on vehicle-mounted 2D LiDAR point clouds data |
| CN108647607A (en) * | 2018-04-28 | 2018-10-12 | 国网湖南省电力有限公司 | Objects recognition method for project of transmitting and converting electricity |
| CN109492699A (en) * | 2018-11-21 | 2019-03-19 | 国网江苏省电力有限公司扬州供电分公司 | Passway for transmitting electricity method for three-dimensional measurement and device |
| CN109492658A (en) * | 2018-09-21 | 2019-03-19 | 北京车和家信息技术有限公司 | A kind of point cloud classifications method and terminal |
| CN109741329A (en) * | 2018-11-27 | 2019-05-10 | 广东工业大学 | A kind of point cloud segmentation method towards electric power corridor scene |
| CN109901618A (en) * | 2019-03-29 | 2019-06-18 | 国网山东省电力公司梁山县供电公司 | Power-line patrolling system and method |
| CN109948414A (en) * | 2018-12-29 | 2019-06-28 | 中国科学院遥感与数字地球研究所 | Electric power corridor scene classification method based on LiDAR point cloud feature |
| CN109948683A (en) * | 2019-03-12 | 2019-06-28 | 百度在线网络技术（北京）有限公司 | Difficulty division methods, device and its relevant device of point cloud data |
| CN110782457A (en) * | 2019-09-06 | 2020-02-11 | 广东工业大学 | Point cloud data segmentation method based on unmanned aerial vehicle power inspection |
| CN111060922A (en) * | 2019-12-11 | 2020-04-24 | 电子科技大学 | Tree point cloud extraction method based on airborne laser radar point cloud spatial distribution characteristics |
| CN111369610A (en) * | 2020-03-05 | 2020-07-03 | 山东交通学院 | Point cloud data gross error positioning and eliminating method based on credibility information |
| CN111553946A (en) * | 2020-04-17 | 2020-08-18 | 中联重科股份有限公司 | Method and device for removing ground point cloud and obstacle detection method and device |
| CN111859772A (en) * | 2020-07-07 | 2020-10-30 | 河南工程学院 | Power line extraction method and system based on cloth simulation algorithm |
| CN112488215A (en) * | 2020-12-04 | 2021-03-12 | 广东电网有限责任公司 | Cable trench point cloud automatic classification method |
| CN112767352A (en) * | 2021-01-19 | 2021-05-07 | 国网通用航空有限公司 | High-voltage line tower automatic fine extraction method for airborne LiDAR power inspection |
| CN113344956A (en) * | 2021-06-21 | 2021-09-03 | 深圳市武测空间信息有限公司 | Ground feature contour extraction and classification method based on unmanned aerial vehicle aerial photography three-dimensional modeling |
| CN113759947A (en) * | 2021-09-10 | 2021-12-07 | 中航空管系统装备有限公司 | Airborne flight obstacle avoidance auxiliary method, device and system based on laser radar |
| CN114462493A (en) * | 2021-12-29 | 2022-05-10 | 浙江大华技术股份有限公司 | Clustering method, clustering device and computer readable storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120128207A1 (en) * | 2010-08-11 | 2012-05-24 | Pasco Corporation | Data analysis device, data analysis method, and program |
| CN102590823A (en) * | 2012-01-06 | 2012-07-18 | 中国测绘科学研究院 | Rapid extraction and reconstruction method for data power line of airborne LIDAR |
| CN102103202B (en) * | 2010-12-01 | 2013-01-09 | 武汉大学 | Semi-supervised classification method for airborne laser radar data fusing images |
| CN102879788A (en) * | 2012-09-04 | 2013-01-16 | 广州建通测绘技术开发有限公司 | Power line extraction method of electric transmission line based on on-board LiDAR data |
| CN103412296A (en) * | 2013-06-28 | 2013-11-27 | 广东电网公司电力科学研究院 | Automatic extraction method of power tower in random laser point cloud data |
-
2015
- 2015-08-31 CN CN201510547324.1A patent/CN105184250B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120128207A1 (en) * | 2010-08-11 | 2012-05-24 | Pasco Corporation | Data analysis device, data analysis method, and program |
| CN102103202B (en) * | 2010-12-01 | 2013-01-09 | 武汉大学 | Semi-supervised classification method for airborne laser radar data fusing images |
| CN102590823A (en) * | 2012-01-06 | 2012-07-18 | 中国测绘科学研究院 | Rapid extraction and reconstruction method for data power line of airborne LIDAR |
| CN102879788A (en) * | 2012-09-04 | 2013-01-16 | 广州建通测绘技术开发有限公司 | Power line extraction method of electric transmission line based on on-board LiDAR data |
| CN103412296A (en) * | 2013-06-28 | 2013-11-27 | 广东电网公司电力科学研究院 | Automatic extraction method of power tower in random laser point cloud data |
Non-Patent Citations (2)
| Title |
|---|
| 董保根: ""机载LiDAR点云与遥感影像融合的地物分类技术研究"", 《中国博士学位论文全文数据库信息科技辑》 * |
| 陈永枫: ""基于机载LiDAR点云数据的建筑物重建技术研究"", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (39)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105701856B (en) * | 2016-01-15 | 2018-10-19 | 厦门大学 | A kind of vegetation extracting method and system |
| CN105701856A (en) * | 2016-01-15 | 2016-06-22 | 厦门大学 | Vegetation extraction method and system |
| CN105701478A (en) * | 2016-02-24 | 2016-06-22 | 腾讯科技（深圳）有限公司 | Method and device for extraction of rod-shaped ground object |
| CN105701478B (en) * | 2016-02-24 | 2019-03-26 | 腾讯科技（深圳）有限公司 | The method and apparatus of rod-shaped Objects extraction |
| CN106199557A (en) * | 2016-06-24 | 2016-12-07 | 南京林业大学 | A kind of airborne laser radar data vegetation extracting method |
| CN106526610A (en) * | 2016-11-04 | 2017-03-22 | 广东电网有限责任公司电力科学研究院 | Power tower automatic positioning method and apparatus based on unmanned aerial vehicle laser point cloud |
| CN106526610B (en) * | 2016-11-04 | 2019-04-09 | 广东电网有限责任公司电力科学研究院 | A kind of pylon automatic positioning method and device based on unmanned plane laser point cloud |
| CN106897686A (en) * | 2017-02-19 | 2017-06-27 | 北京林业大学 | A kind of airborne LIDAR electric inspection process point cloud classifications method |
| CN107273902A (en) * | 2017-05-19 | 2017-10-20 | 中国科学院遥感与数字地球研究所 | A kind of method that electric tower point cloud is automatically extracted from on-board LiDAR data |
| CN107767389A (en) * | 2017-10-27 | 2018-03-06 | 云南电网有限责任公司电力科学研究院 | A kind of power transmission line extraction method based on airborne laser point cloud data |
| CN108564650A (en) * | 2018-01-08 | 2018-09-21 | 南京林业大学 | Shade tree target recognition methods based on vehicle-mounted 2D LiDAR point clouds data |
| CN108564650B (en) * | 2018-01-08 | 2022-03-04 | 南京林业大学 | Lane tree target identification method based on vehicle-mounted 2D LiDAR point cloud data |
| CN108562885B (en) * | 2018-03-29 | 2021-12-31 | 同济大学 | High-voltage transmission line airborne LiDAR point cloud extraction method |
| CN108562885A (en) * | 2018-03-29 | 2018-09-21 | 同济大学 | A kind of ultra-high-tension power transmission line airborne LiDAR point cloud extracting method |
| CN108647607A (en) * | 2018-04-28 | 2018-10-12 | 国网湖南省电力有限公司 | Objects recognition method for project of transmitting and converting electricity |
| CN109492658A (en) * | 2018-09-21 | 2019-03-19 | 北京车和家信息技术有限公司 | A kind of point cloud classifications method and terminal |
| CN109492699A (en) * | 2018-11-21 | 2019-03-19 | 国网江苏省电力有限公司扬州供电分公司 | Passway for transmitting electricity method for three-dimensional measurement and device |
| CN109741329A (en) * | 2018-11-27 | 2019-05-10 | 广东工业大学 | A kind of point cloud segmentation method towards electric power corridor scene |
| CN109948414A (en) * | 2018-12-29 | 2019-06-28 | 中国科学院遥感与数字地球研究所 | Electric power corridor scene classification method based on LiDAR point cloud feature |
| CN109948683A (en) * | 2019-03-12 | 2019-06-28 | 百度在线网络技术（北京）有限公司 | Difficulty division methods, device and its relevant device of point cloud data |
| CN109948683B (en) * | 2019-03-12 | 2021-07-27 | 百度在线网络技术（北京）有限公司 | Method and device for dividing difficulty degree of point cloud data and related equipment thereof |
| CN109901618A (en) * | 2019-03-29 | 2019-06-18 | 国网山东省电力公司梁山县供电公司 | Power-line patrolling system and method |
| CN110782457A (en) * | 2019-09-06 | 2020-02-11 | 广东工业大学 | Point cloud data segmentation method based on unmanned aerial vehicle power inspection |
| CN111060922A (en) * | 2019-12-11 | 2020-04-24 | 电子科技大学 | Tree point cloud extraction method based on airborne laser radar point cloud spatial distribution characteristics |
| CN111060922B (en) * | 2019-12-11 | 2023-04-18 | 电子科技大学 | Tree point cloud extraction method based on airborne laser radar point cloud spatial distribution characteristics |
| CN111369610B (en) * | 2020-03-05 | 2022-09-06 | 山东交通学院 | Point cloud data gross error positioning and eliminating method based on credibility information |
| CN111369610A (en) * | 2020-03-05 | 2020-07-03 | 山东交通学院 | Point cloud data gross error positioning and eliminating method based on credibility information |
| CN111553946A (en) * | 2020-04-17 | 2020-08-18 | 中联重科股份有限公司 | Method and device for removing ground point cloud and obstacle detection method and device |
| CN111553946B (en) * | 2020-04-17 | 2023-04-18 | 中联重科股份有限公司 | Method and device for removing ground point cloud and method and device for detecting obstacle |
| CN111859772A (en) * | 2020-07-07 | 2020-10-30 | 河南工程学院 | Power line extraction method and system based on cloth simulation algorithm |
| CN111859772B (en) * | 2020-07-07 | 2023-11-17 | 河南工程学院 | Power line extraction method and system based on cloth simulation algorithm |
| CN112488215A (en) * | 2020-12-04 | 2021-03-12 | 广东电网有限责任公司 | Cable trench point cloud automatic classification method |
| CN112767352B (en) * | 2021-01-19 | 2023-07-07 | 国网电力空间技术有限公司 | Automatic refined extraction method for high-voltage line tower for airborne LiDAR power inspection |
| CN112767352A (en) * | 2021-01-19 | 2021-05-07 | 国网通用航空有限公司 | High-voltage line tower automatic fine extraction method for airborne LiDAR power inspection |
| CN113344956A (en) * | 2021-06-21 | 2021-09-03 | 深圳市武测空间信息有限公司 | Ground feature contour extraction and classification method based on unmanned aerial vehicle aerial photography three-dimensional modeling |
| CN113759947A (en) * | 2021-09-10 | 2021-12-07 | 中航空管系统装备有限公司 | Airborne flight obstacle avoidance auxiliary method, device and system based on laser radar |
| CN113759947B (en) * | 2021-09-10 | 2023-08-08 | 中航空管系统装备有限公司 | Airborne flight obstacle avoidance assisting method, device and system based on laser radar |
| CN114462493A (en) * | 2021-12-29 | 2022-05-10 | 浙江大华技术股份有限公司 | Clustering method, clustering device and computer readable storage medium |
| CN114462493B (en) * | 2021-12-29 | 2024-05-10 | 浙江大华技术股份有限公司 | Clustering method, clustering device and computer readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105184250B (en) | 2018-06-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105184250A (en) | Ground object classification method based on electric power corridor airborne LiDAR point cloud data | |
| CN103473734B (en) | A kind of electric lines of force based on in-vehicle LiDAR data extracts and approximating method | |
| CN106157309B (en) | A kind of airborne LiDAR ground point cloud filtering method based on virtual seed point | |
| CN108010092A (en) | A kind of city high density area Solar use potential evaluation method based on low altitude photogrammetry | |
| KR101258668B1 (en) | Korea local radar processing system | |
| Kim et al. | 3D classification of power-line scene from airborne laser scanning data using random forests | |
| CN110794413B (en) | Method and system for detecting power line of point cloud data of laser radar segmented by linear voxels | |
| CN113066162B (en) | Urban environment rapid modeling method for electromagnetic calculation | |
| CN109919237B (en) | Point cloud processing method and device | |
| CN103020342A (en) | Method for extracting contour and corner of building from ground LiDAR data | |
| CN104134216B (en) | The laser point cloud autoegistration method described based on 16 dimensional features and system | |
| CN102103202A (en) | Semi-supervised classification method for airborne laser radar data fusing images | |
| CN104050715A (en) | High-precision three-dimensional reconstruction method for power transmission line and corridor | |
| CN107679458B (en) | Method for extracting road marking lines in road color laser point cloud based on K-Means | |
| CN109241978A (en) | The rapid extracting method of planar chip in ground three-dimensional laser point cloud | |
| CN109344812A (en) | A kind of improved single photon point cloud data denoising method based on cluster | |
| CN111323788B (en) | Building change monitoring method and device and computer equipment | |
| CN105550688A (en) | Method and device for point cloud data classification | |
| Hu et al. | A fast and simple method of building detection from LiDAR data based on scan line analysis | |
| CN104597449B (en) | A kind of airborne many scanning weather radar target vertical contour reconstruction methods | |
| CN109492606A (en) | Multispectral vector picture capturing method and system, three dimensional monolithic method and system | |
| CN113963259A (en) | Street view ground object multi-dimensional extraction method and system based on point cloud data | |
| CN111458691B (en) | Building information extraction method and device and computer equipment | |
| CN116071530B (en) | Building roof voxelized segmentation method based on airborne laser point cloud | |
| CN113066161B (en) | Modeling method of urban radio wave propagation model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |