CN110047133A - A kind of train boundary extraction method towards point cloud data - Google Patents
A kind of train boundary extraction method towards point cloud data Download PDFInfo
- Publication number
- CN110047133A CN110047133A CN201910304140.0A CN201910304140A CN110047133A CN 110047133 A CN110047133 A CN 110047133A CN 201910304140 A CN201910304140 A CN 201910304140A CN 110047133 A CN110047133 A CN 110047133A
- Authority
- CN
- China
- Prior art keywords
- point
- train
- point cloud
- boundary
- grid
- 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.)
- Pending
Links
- 238000000605 extraction Methods 0.000 title claims abstract description 19
- 238000000034 method Methods 0.000 claims abstract description 14
- 239000006185 dispersion Substances 0.000 claims abstract description 3
- 239000013598 vector Substances 0.000 claims description 19
- 238000005070 sampling Methods 0.000 claims description 5
- 239000000284 extract Substances 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 238000001514 detection method Methods 0.000 claims description 2
- 238000004919 topotaxy Methods 0.000 claims description 2
- 230000004807 localization Effects 0.000 abstract 1
- 238000005516 engineering process Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- 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
-
- 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
- G06T2207/10032—Satellite or aerial image; Remote sensing
Abstract
This patent proposes a kind of train boundary extraction method towards point cloud data.The method includes the steps: (1) three-dimensional laser scanner acquisition to train initial dispersion point cloud；(2) initialization process of train point cloud filters a cloud and rebuilds；(3) determination of train point cloud boundary and Objective extraction, judgement and acquisition to train point cloud boundary point；(4) generation of train point cloud boundary line is connected to target point cloud boundary point with k nearest neighbor algorithm.The size pose that can effectively know the point cloud data Boundary Extraction of train train does not spend the point cloud considered other than train boundary, suits the remedy to the case, to realize higher efficiency to handling goods ON TRAINS and exact localization operation, more automate.
Description
Technical field
The present invention relates to remote sensing science and technology field, especially a kind of train boundary extraction method towards point cloud data.
Technical background
With the rapid development of laser measuring technique and computer technology, product of the future surface digitizing measuring device is obtained
Popularization and application are arrived.But mainly or manual operation crane either crane is relied on for train handling goods,
It is loaded and unloaded under the commander of relevant staff, positioning is also to complete by the experience of staffing either operator, thus
Cause working efficiency low, is unfavorable for efficient management operation.However there are many developed countries to have been realized in three in foreign countries
The digital measurement device that comes to the surface of dimension laser scanner etc has been used in the handling goods and positioning of train, and industry is realized
High efficiency and automation.
Therefore, China's train handling goods and positioning are improved there are the degree of automation that low, production efficiency is low, operation precision is low
Etc. various disadvantages, just it is particularly important.
Summary of the invention
It is ineffective it is an object of the invention to solve work cumbersome present in train handling goods and positioning
Situation.Selection obtains out the point cloud of train with three-dimensional laser scanner, because operation does not need all environment point clouds, only needs
Cloud boundary information is put to obtain the size pose of train, and then is made a policy, realizes the interaction with ambient enviroment.Point cloud boundary
The important geometrical characteristic of expression curved surface is served not only as, and as the domain for solving curved surface, to the quality for rebuilding surface model
It plays an important role with precision.
In order to achieve the above object, the present invention provides a kind of train boundary extraction method towards point cloud data.Method is logical
Following technical scheme is crossed to realize, the specific steps are as follows:
1) initialization process of train point cloud；
2) determination of train point cloud boundary and Objective extraction；
3) generation of train point cloud boundary line；
Further, specific step is as follows for the initialization process of the point of train described in step 1) cloud:
The original dispersion point cloud of train 1-1) is obtained with three-dimensional laser scanner.Finger gets train from laser scanner
Relative to the three-dimensional coordinate (each pair of point answers a three-dimensional coordinate) of a coordinate origin, which can be selected any solid
Fixed point.Laser scanner is at least greater than the length and width of train to the acquisition length and width of whole point cloud；
Topological relation 1-2) is established with KD tree to cloud.Refer to that the space obtained in train point cloud at random between points is several
What relationship；
1-3) it is filtered according to the actual distribution situation of cloud.Finger removes unrelated to boundary is sought in train point cloud
Point cloud；
1-4) for having situations such as measurement error, outlier, surface hole present in cloud, can be calculated by iteration
Method is rebuild.Refer to the desired profile that train point cloud is obtained by point cloud algorithm for reconstructing.
Further, specific step is as follows for the determination of train point cloud boundary described in step 2) and Objective extraction:
2-1) determine wide W, high H and the ground sampling interval GSD of point cloud characteristic image；
2-2) determine the characteristic value F of each grid (i, j)ij；Assuming that falling in the laser scanning point in (i, j) a grid
Number is nij, the central point of grid (i, j) is P0=(x0,y0, 0), utilize nijThe three-dimensional coordinate weighted calculation grid of a scanning element
The characteristic value F of (i, j)ij。
Point cloud characteristic image 2-3) is generated, by characteristic value FijNaturalization is to 0-255 gray space, to obtain entire scanning area
The point cloud characteristic image in domain；
K closest approach 2-4) is fitted to least square tangent plane, then projects to point in the plane；
Whether uniformly edge feature point 2-5) is judged using the distribution of data point and its k neighborhood, and this distribution is equal
Even property module extracts boundary point by the way of the sum of vector in the field k, when the size and institute's directed quantity of the sum of vector
The ratio for reaching maximum value m in same direction is more than the threshold value Shi Zewei boundary point of a certain setting, is otherwise internal point.
Further, specific step is as follows for the generation of the line of train point cloud boundary described in step 3):
3-1) in order to establish spatial topotaxy between points, need the boundary point of extraction re-establishing KD tree,
To facilitate K nearest neighbor search；
Point cloud boundary line 3-2) is generated using quick sort.Corner dimension sequence is obtained by quick sort, then
It is connected by k nearest neighbor algorithm and obtains train boundary.
Compared with prior art, the invention has the benefit that
1) pass through a cloud characteristic value FijMethod generates the characteristic image of a cloud, chooses from huge point cloud data typical
Characteristic value improves the rapidity and typical accuracy of Boundary Extraction.
2) boundary point is extracted by using the method for the sum of vector in the field k, is cut wherein point is projected to least square
Mode in plane can exclude the interference of lengthy and jumbled data, result can be made with more generation by choosing wherein characteristic feature data
Table.
3) generally speaking, the selection of characteristic feature amount and quick sort obtain corner dimension sequence in the present invention, then
It is connected by k nearest neighbor algorithm and obtains train boundary.Generation to point cloud boundary line is all beneficial to the quick of train boundary acquisition
The requirement of property and typicalness.
Detailed description of the invention
In order to keep the purpose of the present invention, technical solution clearer, the present invention is made below in conjunction with attached drawing further
Detailed description, in which:
Fig. 1 is general frame schematic diagram of the invention；
Fig. 2 is the flow chart that KD tree establishes topological relation in the initialization process of train point cloud；
Fig. 3 is the flow chart of the determination of train point cloud boundary and the generation of Objective extraction midpoint cloud characteristic image；
Fig. 4 is projection point set geometry Characteristics of Distribution flow chart；
Fig. 5 is that train point cloud boundary characteristics extract flow chart；
Specific embodiment
Below in conjunction with attached drawing, a preferred embodiment of the present invention will be described in detail.
Illustrated based on general frame embodiment of the invention as shown in Figure 1, its step are as follows:
Step 1: the point cloud data of train is obtained with three-dimensional laser scanner；
Step 2: KD tree establishes point cloud topological relation；
Step 3: filtering, resurfacing；
Step 4: point cloud characteristic image generates；
Step 5: k closest approach is fitted on least square section and subpoint to plane；
Step 6: train point cloud boundary characteristics point is extracted；
Step 7: train point cloud boundary line generates.
KD tree establishes the flow embodiment signal of topological relation such as in initialization process based on train point cloud of the invention
Shown in Fig. 2, its step are as follows:
Step 1: the intermediate value of each component of three-dimensional point cloud x, y, z is sought.
Step 2: respectively using respective components intermediate value as the 0th layer of KD tree, the discriminant value of layers 1 and 2.Any one
Three-dimensional point Pi, F layers of KD tree, then have n=Fmod3, which of x, y, z component determined by n value.As n=0,
Take x-component；As n=1, y-component is taken；As n=2, z-component is taken.
Step 3: comparing division discriminant value, is divided into left branch less than discriminant value, greater than assigning to right branch.Work as step 2
When calculating the value for obtaining component less than or equal to the discriminant value compared, then this three-dimensional point is just divided to left branch；Greater than discriminant value
When be just divided to right branch.
Step 4: until all the points are inserted into KD tree, topological relation building finishes circulate operation.
Based on the flow implementation of the determination of train point cloud boundary and the generation of Objective extraction midpoint cloud characteristic image of the invention
Meaning is illustrated as shown in figure 3, its step are as follows:
Step 1: wide W, high H and the ground sampling interval GSD of point cloud characteristic image are determined.It, can be with by point cloud data
Learn that minimax X, Y, the Z coordinate of scanning area are respectively as follows: Xmin、Ymin、Zmin、Xmax、Ymax、Zmax, the customized ground of user
Sampling interval GSD, then have:
Step 2: the characteristic value F of each grid (i, j) is determinedij.Assuming that falling in the laser scanning in (i, j) a grid
Point number is nij, the central point of grid (i, j) is P0 H=(x0 H,y0 H, 0), utilize nijThe three-dimensional coordinate of a scanning element weights meter
Calculate the characteristic value F of grid (i, j)ij.In order to determine fall in cell (i, j) it is each (such as k-th point, 0 < k≤nij) power
Value Wijk.By the characteristic value F of gridijCalculating is divided into two parts: first part is by between all the points in grid and grid central point
XOY plane distance Dij kIt determines；Second part is by the elevation difference H between minimum point in all the points in grid and gridij kCertainly
It is fixed.
In formula, Wijk XY、Wijk HThe power of respectively scanning element k weight and scanning element k elevation at a distance from grid points center
Value.hmin(ij)、hmax(ij)Minimum elevation and highest elevation respectively in grid (i, j), Zmax、ZminIt is entire scanning area respectively
The highest elevation in domain and minimum elevation.If ZijkFor the height value of k-th of point P (i, j, k) in grid (i, j), then Hij k=
Zij k-Zmin。Wijk XYReflect contribution of the plan range of discrete point and grid central point to grid central point characteristic value.Wijk HInstead
Contribution of the elevation of discrete point in grid to grid central point characteristic value is reflected.
By setting different α, β value, the characteristic value of grid (i, j) can be calculated using IDW interpolating method, is described
Are as follows:
Step 3: point cloud characteristic image is generated.By grid feature value FijGrid can be obtained to 0-255 gray space in naturalization
Corresponding cloud characteristic image.
Based on it is of the invention be projection point set geometry Characteristics of Distribution flow embodiment signal as shown in figure 4, its step
It is as follows:
Step 1: by point set X to shown in the projection of its tangent plane, projection point set is X={ (xi,yi,zi| i=0,1 ...,
k)}.With the subpoint P of current pointiFor starting point, NiFor terminal definition vector PiNj, a vector P is taken wherein appointingiNj-1, ask its with
The cross product v and other vectors and vector P of tangent plane normal vectoriNj-1With the angle α of vi, βjIf βj>=90 °, then αj=360 °-
αj。
Step 2: utilize quick sort by angle αjIt sorts from large to small, then the angle between adjacent vector may be expressed as:
Flow embodiment is extracted based on train point cloud boundary characteristics of the invention to illustrate as shown in figure 5, its step are as follows:
Step 1: based on the maximum angular δ between adjacent vectormaxWhen greater than angle threshold ε, it can determine whether that current point is boundary point.
Step 2: for choosing a point in the boundary point that has judged as seed point, then along a fixed-direction
(clockwise either counterclockwise) passes through nearest another boundary point B of detection range seed pointi, search again for distance Bi
The boundary B that do not search for recently andi+1..., until searching seed point location again.Thus successfully track out institute's mesh to be separated
Mark boundary point.
Claims (4)
1. a kind of train boundary extraction method towards point cloud data, which comprises the following steps:
Step 1: the initialization process of train point cloud；
Step 2: the determination of train point cloud boundary and Objective extraction；
Step 3: the generation of train point cloud boundary line.
2. a kind of train boundary extraction method towards point cloud data according to claim 1, it is characterised in that: the step
Rapid one the following steps are included:
21: the original dispersion point cloud of train is obtained with three-dimensional laser scanner.It is opposite that finger gets train from laser scanner
In the three-dimensional coordinate (each pair of point answers a three-dimensional coordinate) of a coordinate origin, which can be selected any fixation
Point.Laser scanner is at least greater than the length and width of train to the acquisition length and width of whole point cloud.
22: topological relation is established with KD tree to cloud.Refer to that the space geometry obtained in train point cloud at random between points closes
System.
221: seeking the intermediate value of each component of three-dimensional point cloud x, y, z.
222: respectively using respective components intermediate value as the 0th layer of KD tree, the discriminant value of layers 1 and 2.Any one three-dimensional point Pi,
F layers of KD tree, then have n=Fmod3, and which of x, y, z component is determined by n value.As n=0, x-component is taken；When
When n=1, y-component is taken；As n=2, z-component is taken.
223: comparing division discriminant value, be divided into left branch less than discriminant value, greater than assigning to right branch.Component is obtained when calculating
When value is less than or equal to the discriminant value compared, then this three-dimensional point is just divided to left branch；Right point is just divided to when greater than discriminant value
Branch.
224: until all the points are inserted into KD tree, topological relation building finishes circulate operation.
23: being filtered according to the actual distribution situation of cloud.Finger removes in train point cloud to seeking the unrelated point in boundary
Cloud.
24: for having situations such as measurement error, outlier, surface hole present in cloud, an iterative algorithm weight can be passed through
It builds.Refer to the desired profile that train point cloud is obtained by point cloud algorithm for reconstructing.
3. a kind of according to claim 1, train boundary extraction method towards point cloud data, it is characterised in that: described
Step 2 the following steps are included:
31: determining wide W, high H and the ground sampling interval GSD of point cloud characteristic image.By point cloud data, scanning can be learnt
Minimax X, Y, the Z coordinate in region are respectively as follows: Xmin、Ymin、Zmin、Xmax、Ymax、Zmax, user's customized ground sampling interval
GSD then has:
32: determining the characteristic value F of each grid (i, j)ij；Assuming that the laser scanning point number fallen in (i, j) a grid is
nij, the central point of grid (i, j) is P0=(x0,y0, 0), utilize nijThe three-dimensional coordinate weighted calculation grid (i, j) of a scanning element
Characteristic value Fij.In order to determine fall in cell (i, j) it is each (such as k-th point, 0 < k≤nij) weight Wijk.By grid
Characteristic value FijCalculating is divided into two parts: first part by the XOY plane between all the points in grid and grid central point away from
From Dij kIt determines；Second part is by the elevation difference H between minimum point in all the points in grid and gridij kIt determines.
In formula, Wijk XY、Wijk HRespectively scanning element k weight of weight and scanning element k elevation at a distance from grid points center.
hmin(ij)、hmax(ij)Minimum elevation and highest elevation respectively in grid (i, j), Zmax、ZminIt is entire scanning area respectively
Highest elevation and minimum elevation.If ZijkFor the height value of k-th of point P (i, j, k) in grid (i, j), then Hij k=Zij k-
Zmin。Wijk XYReflect contribution of the plan range of discrete point and grid central point to grid central point characteristic value.Wijk HIt reflects
Contribution of the elevation of discrete point to grid central point characteristic value in grid.
By setting different α, β value, the characteristic value of grid (i, j) can be calculated using IDW interpolating method, is described are as follows:
33: point cloud characteristic image is generated, by characteristic value FijNaturalization is to 0-255 gray space, to obtain entire scanning area
Point cloud characteristic image.
34: k closest approach being fitted to least square tangent plane, then projects to point in the plane.By point set X to its tangent plane
Shown in projection, projection point set is X={ (xi,yi,zi| i=0,1 ..., k).With the subpoint P of current pointiFor starting point, Ni
For terminal definition vector PiNj, a vector P is taken wherein appointingiNj-1, ask its cross product v and other vectors with tangent plane normal vector
With vector PiNj-1With the angle α of vi, βjIf βj>=90 °, then αj=360 ° of-αj。
35: whether uniformly edge feature point is judged using the distribution of data point and its k neighborhood, and the uniformity metric of this distribution
Amount standard extracts boundary point by the way of the sum of vector in the field k, when the sum of vector size with institute directed quantity same
The ratio for reaching maximum value m when direction is more than the threshold value Shi Zewei boundary point of a certain setting, is otherwise internal point.
4. a kind of according to claim 1, train boundary extraction method towards point cloud data, it is characterised in that: described
Step 3 the following steps are included:
41: in order to establish spatial topotaxy between points, needing the boundary point of extraction re-establishing KD tree, with convenient
K nearest neighbor search.
42: point cloud boundary line is generated using quick sort.Corner dimension sequence is obtained by quick sort, then passes through k
Nearest neighbor algorithm, which is connected, obtains train boundary.
421: by angle αjIt sorts from large to small, then the angle between adjacent vector may be expressed as:
422: based on the maximum angular δ between adjacent vectormaxWhen greater than angle threshold ε, it can determine whether that current point is boundary point.
423: (clockwise then along a fixed-direction for choosing a point in the boundary point that has judged as seed point
Either counterclockwise), another boundary point B nearest by detection range seed pointi, search again for distance BiRecently and
The boundary B that do not search fori+1..., until searching seed point location again.Thus successfully track out institute's object boundary point to be separated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910304140.0A CN110047133A (en) | 2019-04-16 | 2019-04-16 | A kind of train boundary extraction method towards point cloud data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910304140.0A CN110047133A (en) | 2019-04-16 | 2019-04-16 | A kind of train boundary extraction method towards point cloud data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110047133A true CN110047133A (en) | 2019-07-23 |
Family
ID=67277360
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910304140.0A Pending CN110047133A (en) | 2019-04-16 | 2019-04-16 | A kind of train boundary extraction method towards point cloud data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110047133A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111210501A (en) * | 2019-10-12 | 2020-05-29 | 香港理工大学深圳研究院 | Indoor modeling method and device and terminal equipment |
| CN111815611A (en) * | 2020-07-14 | 2020-10-23 | 南京航空航天大学苏州研究院 | Round hole feature extraction method for rivet hole measurement point cloud data |
| CN112068128A (en) * | 2020-09-19 | 2020-12-11 | 重庆大学 | Method for processing data and acquiring pose of linear radar in straight road scene |
| CN112819952A (en) * | 2021-02-24 | 2021-05-18 | 中冶南方工程技术有限公司 | Method for determining optimal material taking point of scrap steel yard disk crane |
| CN113034539A (en) * | 2019-12-25 | 2021-06-25 | 华为技术有限公司 | Method and device for determining boundary frame of point cloud |
| CN113409347A (en) * | 2021-08-19 | 2021-09-17 | 深圳市信润富联数字科技有限公司 | Method and device for extracting point cloud boundary, storage medium and electronic equipment |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101726255A (en) * | 2008-10-24 | 2010-06-09 | 中国科学院光电研究院 | Method for extracting interesting buildings from three-dimensional laser point cloud data |
| CN103247045A (en) * | 2013-04-18 | 2013-08-14 | 上海交通大学 | Method of obtaining artificial scene main directions and image edges from multiple views |
| CN104200212A (en) * | 2014-06-25 | 2014-12-10 | 西安煤航信息产业有限公司 | Building outer boundary line extraction method based on onboard LiDAR (Light Detection and Ranging) data |
| CN106970581A (en) * | 2017-04-30 | 2017-07-21 | 中南大学 | A kind of train pantograph real-time intelligent monitoring method and system based on the three-dimensional full visual angle of unmanned aerial vehicle group |
| CN107123164A (en) * | 2017-03-14 | 2017-09-01 | 华南理工大学 | Keep the three-dimensional rebuilding method and system of sharp features |
| CN108198244A (en) * | 2017-12-20 | 2018-06-22 | 中国农业大学 | A kind of Apple Leaves point cloud compressing method and device |
| CN109410342A (en) * | 2018-09-28 | 2019-03-01 | 昆明理工大学 | A kind of point cloud compressing method retaining boundary point |
-
2019
- 2019-04-16 CN CN201910304140.0A patent/CN110047133A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101726255A (en) * | 2008-10-24 | 2010-06-09 | 中国科学院光电研究院 | Method for extracting interesting buildings from three-dimensional laser point cloud data |
| CN103247045A (en) * | 2013-04-18 | 2013-08-14 | 上海交通大学 | Method of obtaining artificial scene main directions and image edges from multiple views |
| CN104200212A (en) * | 2014-06-25 | 2014-12-10 | 西安煤航信息产业有限公司 | Building outer boundary line extraction method based on onboard LiDAR (Light Detection and Ranging) data |
| CN107123164A (en) * | 2017-03-14 | 2017-09-01 | 华南理工大学 | Keep the three-dimensional rebuilding method and system of sharp features |
| CN106970581A (en) * | 2017-04-30 | 2017-07-21 | 中南大学 | A kind of train pantograph real-time intelligent monitoring method and system based on the three-dimensional full visual angle of unmanned aerial vehicle group |
| CN108198244A (en) * | 2017-12-20 | 2018-06-22 | 中国农业大学 | A kind of Apple Leaves point cloud compressing method and device |
| CN109410342A (en) * | 2018-09-28 | 2019-03-01 | 昆明理工大学 | A kind of point cloud compressing method retaining boundary point |
Non-Patent Citations (4)
| Title |
|---|
| 刘宇 等: "基于倾斜影像点云的建筑物提取算法研究", 《河南科技》 * |
| 孙伟 等: "《嵌入式图像处理系统工程案例精选-DaVinci系统设计与实现》", 30 April 2017 * |
| 费爱国 等: "《多传感器数据融合—模型、方法与实践》", 30 November 2006 * |
| 邱莉榕 等: "《算法设计与优化》", 31 December 2016 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111210501A (en) * | 2019-10-12 | 2020-05-29 | 香港理工大学深圳研究院 | Indoor modeling method and device and terminal equipment |
| CN113034539A (en) * | 2019-12-25 | 2021-06-25 | 华为技术有限公司 | Method and device for determining boundary frame of point cloud |
| WO2021129483A1 (en) * | 2019-12-25 | 2021-07-01 | 华为技术有限公司 | Method for determining point cloud bounding box, and apparatus |
| CN111815611A (en) * | 2020-07-14 | 2020-10-23 | 南京航空航天大学苏州研究院 | Round hole feature extraction method for rivet hole measurement point cloud data |
| CN111815611B (en) * | 2020-07-14 | 2024-03-22 | 南京航空航天大学 | Round hole feature extraction method for rivet hole measurement point cloud data |
| CN112068128A (en) * | 2020-09-19 | 2020-12-11 | 重庆大学 | Method for processing data and acquiring pose of linear radar in straight road scene |
| CN112819952A (en) * | 2021-02-24 | 2021-05-18 | 中冶南方工程技术有限公司 | Method for determining optimal material taking point of scrap steel yard disk crane |
| CN112819952B (en) * | 2021-02-24 | 2022-05-13 | 中冶南方工程技术有限公司 | Method for determining optimal material taking point of scrap steel yard disk crane |
| CN113409347A (en) * | 2021-08-19 | 2021-09-17 | 深圳市信润富联数字科技有限公司 | Method and device for extracting point cloud boundary, storage medium and electronic equipment |
| CN113409347B (en) * | 2021-08-19 | 2021-12-21 | 深圳市信润富联数字科技有限公司 | Method and device for extracting point cloud boundary, storage medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110047133A (en) | A kind of train boundary extraction method towards point cloud data | |
| CN107038717B (en) | A method of 3D point cloud registration error is automatically analyzed based on three-dimensional grid | |
| CN106023298B (en) | Point cloud Rigid Registration method based on local Poisson curve reestablishing | |
| CN107063228B (en) | Target attitude calculation method based on binocular vision | |
| CN108133458A (en) | A kind of method for automatically split-jointing based on target object spatial point cloud feature | |
| CN109559338B (en) | Three-dimensional point cloud registration method based on weighted principal component analysis method and M estimation | |
| CN107292234B (en) | Indoor scene layout estimation method based on information edge and multi-modal features | |
| CN110081888B (en) | Indoor positioning algorithm based on credibility-based inertial navigation and geomagnetic fusion | |
| CN110807781B (en) | Point cloud simplifying method for retaining details and boundary characteristics | |
| CN108171780A (en) | A kind of method that indoor true three-dimension map is built based on laser radar | |
| JP4205760B1 (en) | Image matching method, program and application apparatus | |
| CN112347550B (en) | Coupling type indoor three-dimensional semantic graph building and modeling method | |
| CN106373118A (en) | A complex curved surface part point cloud reduction method capable of effectively keeping boundary and local features | |
| CN103136525B (en) | A kind of special-shaped Extended target high-precision locating method utilizing Generalized Hough Transform | |
| CN105654483A (en) | Three-dimensional point cloud full-automatic registration method | |
| Li et al. | Automatic evaluation of machining allowance of precision castings based on plane features from 3D point cloud | |
| CN108550166B (en) | Spatial target image matching method | |
| Nevado et al. | Obtaining 3D models of indoor environments with a mobile robot by estimating local surface directions | |
| CN111932669A (en) | Deformation monitoring method based on slope rock mass characteristic object | |
| CN112381862B (en) | Full-automatic registration method and device for CAD (computer-aided design) model and triangular mesh | |
| CN110942077B (en) | Feature line extraction method based on weight local change degree and L1 median optimization | |
| CN109448040A (en) | A kind of machinery production manufacture displaying auxiliary system | |
| CN110060344A (en) | A kind of prefabricated components overall dimensions reverse modeling method based on point cloud data | |
| CN109345571B (en) | Point cloud registration method based on extended Gaussian image | |
| Kaushik et al. | Accelerated patch-based planar clustering of noisy range images in indoor environments for robot mapping |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190723 |
|
| RJ01 | Rejection of invention patent application after publication |