CN104574303A - Airborne LiDAR point cloud ground filtering method based on spatial clustering - Google Patents

Abstract

The invention discloses an airborne LiDAR point cloud ground filtering method based on spatial clustering. The method comprises steps as follows: airborne LiDAR original point cloud data are preprocessed, and abnormal values are removed; multiple-echo information in the preprocessed point cloud data is analyzed, and single-echo information and final-echo information in the information are selected as processed point cloud data; a plane model of the processed point cloud data is adopted as a segmentation stop condition for point cloud segmentation; the segmented point cloud data are subjected to spatial clustering with a dual-distance method, and coarse classification of the point cloud data is finished; an initial triangulation network is built according to ground clustering blocks obtained after coarse classification; according to the initial triangulation network, a preset threshold value is taken as a judgment condition to identify whether other to-be-judged clustering blocks are ground blocks or non-ground blocks; a final triangulation network is generated through iterative interpolation, and net-constructing data are ground point data and are output. The overall filtering effect is more reliable.

Description

Based on the airborne LiDAR point cloud ground filtering method of space clustering

Technical field

The present invention relates to airborne laser radar data process field technical field, particularly relate to a kind of airborne LiDAR point cloud ground filtering method based on space clustering.

Background technology

Along with the fast development of remote sensing system, airborne laser detection and range finding (light detection and ranging, LiDAR) more and more become a kind of core technology of current quick obtaining 3-dimensional digital ground model.LiDAR utilizes GPS (Global Position System, and inertial measuring unit (Inertial Measurement Unit GPS), IMU) airborne lidar is carried out, measured data are digital surface model (Digital Surface Model, DSM) discrete point cloud data, containing space three-dimensional information and laser intensity information in data, sorting technique is adopted to remove buildings again in digital surface model, artificiality, the measuring points such as cover plant, digital elevation model (Digital Elevation Model can be obtained, DEM), and obtain the height of mulching material simultaneously.Namely ground filtering be the process distinguishing ground point and non-ground points from discrete cloud data, is one of steps necessary of on-board LiDAR data aftertreatment.

So far, many documents are proposed the ground filtering method about airborne LiDAR point cloud data, wherein based on cluster segmentation filtering method it is considered that relation between the set of similar point, and not only rely on textural difference between points as topographic structure criterion, therefore judge to identify at terrain and its features and have more rationally, filter result is more reliable.But in cluster process, carry out cluster iff according to the height between atural object or topological relation, often cause cluster unreasonable, or lose effective information, be therefore necessary that the characteristic of addition point cloud carries out cluster.

Summary of the invention

Technical matters to be solved by this invention is, a kind of airborne LiDAR point cloud ground filtering method based on space clustering is provided, for ground non-continuous event, solve the restriction of single threshold value to filter effect, and complicated terrain feature can be effectively retained while filtering terrestrial object information, make overall filter effect more reliable.

In order to solve the problems of the technologies described above, the invention provides a kind of airborne LiDAR point cloud ground filtering method based on space clustering, comprising:

Pre-service is carried out to airborne LiDAR original point cloud data, thus rejecting abnormalities value;

Multiecho information in pretreated cloud data is analyzed, and the single echo information chosen wherein and last echo information are as process cloud data;

The areal model of described process cloud data is adopted to split for segmentation stop condition carries out a cloud;

Adopt dual distance method to carry out space clustering to the cloud data after segmentation, complete cloud data rough sort;

Ground cluster block according to obtaining after rough sort builds the initial triangulation network;

According to the described initial triangulation network, be Rule of judgment with predetermined threshold value, identify that other waits to judge that cluster block is ground block or non-ground block;

Iterative interpolation generates the final triangulation network, and export network forming data, described network forming data are ground point data.

Wherein, described pre-service is carried out to airborne LiDAR original point cloud data, thus rejecting abnormalities value, specifically comprise:

Statistical analysis technique is adopted to carry out denoising to airborne LiDAR original point cloud data, thus rejecting abnormalities value.

Wherein, the areal model of the described process cloud data of described employing carries out a cloud segmentation for segmentation stop condition, specifically comprises:

Octree index is set up according to described process cloud data;

Adopt described Octree index as searching method used during the described process cloud data of extraction, thus by the areal model of described process cloud data for segmentation stop condition carries out a cloud segmentation.

Wherein, in the areal model of described process cloud data, arbitrfary point P in described process cloud data _i(x _i, y _i, z _i) to the vertical range of plane be: V _i=| ax _i+ by _i+ cz _i+ d|, wherein, the normal vector parameter that (a, b, c) is plane, a, b, c, d are the parameter of plane respectively.

Wherein, the dual distance method of described employing carries out space clustering to the cloud data after segmentation, completes cloud data rough sort, specifically comprises:

Adopt the dual distance method comprising space length and attributive distance to carry out space clustering to the cloud data after segmentation, complete cloud data rough sort, wherein,

F _ito f _jspace length: $D_{\mathrm{geo}}(f_i,f_j)=\sqrt{{(x_i-x_j)}^2+{(y_i-y_j)}^2+{(z_i-z_j)}^2}$

F _ito f _jattributive distance be: $D_{\mathrm{art}}(f_i,f_j)={\left(\underset{k=1}{\overset{m}{\mathrm{\Σ}}}\frac{|A_{\mathrm{ik}}-A_{\mathrm{jk}}|}{D_{k\max }}\right)}^{\frac{1}{m}}$

Wherein, f _i, f _j(1≤i, j≤n) is three dimensions key element collection F={f ₁, f ₂..., f _n| the key element in n>=2}, m is non-spatial attributes dimension, A _ikfor kth dimension attribute, D _kmaxfor the difference of a kth attribute minimax attribute, i.e. D _kmax=A _kmax-A _kmin.

Wherein, described attribute information is a cloud strength information and the normal information according to the fit Plane of point set in least square method calculation level cloud neighborhood.

Wherein, the described ground cluster block according to obtaining after rough sort builds the initial triangulation network, specifically comprises:

Maximum ground cluster block according to obtaining after rough sort builds the initial ground triangulation network as parameter request.

Implement the present invention, there is following beneficial effect: the present invention is split to construct octree structure, take areal model as segmentation stop condition, effectively can avoid the over-segmentation phenomenon caused based on partial points cloud and region growing methods; Carry out space clustering by the method for dual distance, account for the attribute information of a cloud, so both can solve the restriction of single threshold value to filter effect, avoid cluster un-reasonable phenomenon, and when processing discontinuous ground, can not only object point discretely, and can effectively keep complicated terrain feature.

Accompanying drawing explanation

In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, be briefly described to the accompanying drawing used required in embodiment or description of the prior art below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.

Fig. 1 is the schematic flow sheet of the airborne LiDAR point cloud ground filtering method based on space clustering that the embodiment of the present invention provides.

Fig. 2 is wait judging the triangulation network schematic diagram corresponding when judging point is located in cluster block.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present invention, be clearly and completely described the technical scheme in the embodiment of the present invention, obviously, described embodiment is only the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making the every other embodiment obtained under creative work prerequisite, belong to the scope of protection of the invention.

Fig. 1 is the schematic flow sheet of the airborne LiDAR point cloud ground filtering method based on space clustering that the embodiment of the present invention provides, and as shown in Figure 1, comprises step:

S101, pre-service is carried out to airborne LiDAR original point cloud data, thus rejecting abnormalities value.

Concrete, S101 comprises step:

Statistical analysis technique is adopted to carry out denoising to airborne LiDAR original point cloud data, thus rejecting abnormalities value.

S102, the multiecho information in pretreated cloud data to be analyzed, and the single echo information chosen wherein and last echo information are as process cloud data.

Wherein, in multiecho information, no matter forest land or urban area, single echo and last echo not only can from ground, but also can from ground object target, echo and centre time echo derive from ground object target, certainly as vegetation, buildings edge etc. first.The Different Ground feature according to the multiecho message reflection of a cloud, echo and centre time echo do not participate in follow-up filtering operation first, choose single echo information wherein and last echo information as process cloud data.

S103, the areal model of described process cloud data is adopted to carry out the segmentation of a some cloud for segmentation stop condition.

Concrete, S103 comprises step:

Octree index is set up according to described process cloud data;

Adopt described Octree index as searching method used during the described process cloud data of extraction, thus by the areal model of described process cloud data for segmentation stop condition carries out a cloud segmentation.

Wherein, in the areal model of described process cloud data, arbitrfary point P in described process cloud data _i(x _i, y _i, z _i) to the vertical range of plane be: V _i=| ax _i+ by _i+ cz _i+ d|, wherein, the normal vector parameter that (a, b, c) is plane, a, b, c, d are the parameter of plane respectively.

S104, adopt dual distance method to segmentation after cloud data carry out space clustering, complete cloud data rough sort.

Concrete, S104 specifically comprises step:

Adopt the dual distance method comprising space length and attributive distance to carry out space clustering to the cloud data after segmentation, complete cloud data rough sort, wherein,

F _ito f _jspace length: $D_{\mathrm{geo}}(f_i,f_j)=\sqrt{{(x_i-x_j)}^2+{(y_i-y_j)}^2+{(z_i-z_j)}^2}$

F _ito f _jattributive distance be: $D_{\mathrm{art}}(f_i,f_j)={\left(\underset{k=1}{\overset{m}{\mathrm{\Σ}}}\frac{|A_{\mathrm{ik}}-A_{\mathrm{jk}}|}{D_{k\max }}\right)}^{\frac{1}{m}}$

Wherein, f _i, f _j(1≤i, j≤n) is three dimensions key element collection F={f ₁, f ₂..., f _n| the key element in n>=2}, m is non-spatial attributes dimension, A _ikfor kth dimension attribute, D _kmaxfor the difference of a kth attribute minimax attribute, i.e. D _kmax=A _kmax-A _kmin.Wherein, described attribute information is a cloud strength information and the normal information according to the fit Plane of point set in least square method calculation level cloud neighborhood.

S105, build the initial triangulation network according to the ground cluster block obtained after rough sort.

Concrete, S105 comprises step: the maximum ground cluster block according to obtaining after rough sort builds the initial ground triangulation network as parameter request.Namely the some cloud in maximum ground cluster block is selected to generate the initial triangulation network as Seed Points.

S106, according to the described initial triangulation network, be Rule of judgment with predetermined threshold value, identify that other waits to judge that cluster block is ground block or non-ground block.

Concrete, wait judging the triangulation network corresponding when judging point is located in cluster block as shown in Figure 2, treat that judging point is d to the triangular facet vertical range that Seed Points generates, treat that the subpoint of judging point and a direction angle of triangular apex are α, if d and α is less than predetermined threshold value, then definition treats that judging point is ground point, and together with treating that the some cloud mass of judging point is judged to be ground block, otherwise, reject and treat judging point and corresponding cluster block, be judged to be that non-ground is fast.

S107, iterative interpolation generate the final triangulation network, and export network forming data, described network forming data are ground point data.

In order to quantitative test filtering method effect of the present invention, the filtering method adopting ISPRS group to issue compares evaluation method in report, and carries out quantitative comparison with the total error of classical filter algorithm in the world:

As can be seen from the above table, the present invention is directed to the discontinuous landform in ground, can be easy to extract ground, keep the feature of landform well, and carry out filtering based on a certain fixed threshold, even if having indivedual local by mis-classification, but the foundation of the progressive triangulation network in block-based ground can reduce classification error, total error is minimum, makes overall filter effect reach best.

Implement the present invention, there is following beneficial effect: the present invention is split to construct octree structure, take areal model as segmentation stop condition, effectively can avoid the over-segmentation phenomenon caused based on partial points cloud and region growing methods; Carry out space clustering by the method for dual distance, account for the attribute information of a cloud, so both can solve the restriction of single threshold value to filter effect, avoid cluster un-reasonable phenomenon, and when processing discontinuous ground, can not only object point discretely, and can effectively keep complicated terrain feature.

It should be noted that, in this article, term " comprises ", " comprising " or its any other variant are intended to contain comprising of nonexcludability, thus make to comprise the process of a series of key element, method, article or device and not only comprise those key elements, but also comprise other key elements clearly do not listed, or also comprise by the intrinsic key element of this process, method, article or device.When not more restrictions, the key element limited by statement " comprising ... ", and be not precluded within process, method, article or the device comprising this key element and also there is other identical element.

Professional can also recognize further, in conjunction with unit and the algorithm steps of each example of embodiment disclosed herein description, can realize with electronic hardware, computer software or the combination of the two, in order to the interchangeability of hardware and software is clearly described, generally describe composition and the step of each example in the above description according to function.These functions perform with hardware or software mode actually, depend on application-specific and the design constraint of technical scheme.Professional and technical personnel can use distinct methods to realize described function to each specifically should being used for, but this realization should not thought and exceeds scope of the present invention.

The software module that the method described in conjunction with embodiment disclosed herein or the step of algorithm can directly use hardware, processor to perform, or the combination of the two is implemented.Software module can be placed in the storage medium of other form any known in random access memory (RAM), internal memory, ROM (read-only memory) (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technical field.

To the above-mentioned explanation of the disclosed embodiments, professional and technical personnel in the field are realized or uses the present invention.To be apparent for those skilled in the art to the multiple amendment of these embodiments, General Principle as defined herein can without departing from the spirit or scope of the present invention, realize in other embodiments.Therefore, the present invention can not be restricted to these embodiments shown in this article, but will meet the widest scope consistent with principle disclosed herein and features of novelty.

Claims (7)

1., based on an airborne LiDAR point cloud ground filtering method for space clustering, it is characterized in that, comprising:

Pre-service is carried out to airborne LiDAR original point cloud data, thus rejecting abnormalities value;

Multiecho information in pretreated cloud data is analyzed, and the single echo information chosen wherein and last echo information are as process cloud data;

The areal model of described process cloud data is adopted to split for segmentation stop condition carries out a cloud;

Adopt dual distance method to carry out space clustering to the cloud data after segmentation, complete cloud data rough sort;

Ground cluster block according to obtaining after rough sort builds the initial triangulation network;

According to the described initial triangulation network, be Rule of judgment with predetermined threshold value, identify that other waits to judge that cluster block is ground block or non-ground block;

Iterative interpolation generates the final triangulation network, and export network forming data, described network forming data are ground point data.

2. as claimed in claim 1 based on the airborne LiDAR point cloud ground filtering method of space clustering, it is characterized in that, described pre-service is carried out to airborne LiDAR original point cloud data, thus rejecting abnormalities value, specifically comprise:

Statistical analysis technique is adopted to carry out denoising to airborne LiDAR original point cloud data, thus rejecting abnormalities value.

3. as claimed in claim 1 based on the airborne LiDAR point cloud ground filtering method of space clustering, it is characterized in that, the areal model of the described process cloud data of described employing carries out a cloud segmentation for segmentation stop condition, specifically comprises:

Octree object is set up according to described process cloud data;

Adopt described Octree object as searching method used during the described process cloud data of extraction, thus by the areal model of described process cloud data for segmentation stop condition carries out a cloud segmentation.

4., as claimed in claim 3 based on the airborne LiDAR point cloud ground filtering method of space clustering, it is characterized in that, in the areal model of described process cloud data, arbitrfary point P in described process cloud data _i(x _i, y _i, z _i) to the vertical range of plane be: V _i=| ax _i+ by _i+ cz _i+ d|, wherein, the normal vector parameter that (a, b, c) is plane, a, b, c, d are the parameter of plane respectively.

5. as claimed in claim 1 based on the airborne LiDAR point cloud ground filtering method of space clustering, it is characterized in that, the dual distance method of described employing carries out space clustering to the cloud data after segmentation, completes cloud data rough sort, specifically comprises:

Adopt the dual distance method comprising space length and attributive distance to carry out space clustering to the cloud data after segmentation, complete cloud data rough sort, wherein,

F _ito f _jspace length: $D_{\mathrm{geo}}(f_i,f_j)=\sqrt{{(x_i-x_j)}^2+{(y_i-y_j)}^2+{(z_i-z_j)}^2}$

F _ito f _jattributive distance be: $D_{\mathrm{art}}(f_i,f_j)={\left(\underset{k=1}{\overset{m}{\mathrm{\Σ}}}\frac{|A_{\mathrm{ik}}-A_{\mathrm{jk}}|}{D_{k\max }}\right)}^{\frac{1}{m}}$

Wherein, f _i, f _j(1≤i, j≤n) is three dimensions key element collection F={f ₁, f ₂..., f _n| the key element in n>=2}, m is non-spatial attributes dimension, A _ikfor kth dimension attribute, D _kmaxfor the difference of a kth attribute minimax attribute, i.e. D _kmax=A _kmax-A _kmin.

6., as claimed in claim 5 based on the airborne LiDAR point cloud ground filtering method of space clustering, it is characterized in that, described attribute information is a cloud strength information and the normal information according to the fit Plane of point set in least square method calculation level cloud neighborhood.

7., as claimed in claim 1 based on the airborne LiDAR point cloud ground filtering method of space clustering, it is characterized in that, the described ground cluster block according to obtaining after rough sort builds the initial triangulation network, specifically comprises:

Maximum ground cluster block according to obtaining after rough sort builds the initial ground triangulation network as parameter request.