CN110111378A - A kind of point cloud registering optimization method and device based on indoor three-dimensional data - Google Patents
A kind of point cloud registering optimization method and device based on indoor three-dimensional data Download PDFInfo
- Publication number
- CN110111378A CN110111378A CN201910270158.3A CN201910270158A CN110111378A CN 110111378 A CN110111378 A CN 110111378A CN 201910270158 A CN201910270158 A CN 201910270158A CN 110111378 A CN110111378 A CN 110111378A
- Authority
- CN
- China
- Prior art keywords
- point
- ground
- ladder
- height
- outer ginseng
- 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
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000005457 optimization Methods 0.000 title claims abstract description 15
- 241000208340 Araliaceae Species 0.000 claims abstract description 92
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims abstract description 92
- 235000003140 Panax quinquefolius Nutrition 0.000 claims abstract description 92
- 235000008434 ginseng Nutrition 0.000 claims abstract description 92
- 239000011159 matrix material Substances 0.000 claims abstract description 89
- 238000004590 computer program Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 16
- 230000008569 process Effects 0.000 description 6
- 230000000694 effects Effects 0.000 description 5
- 241001269238 Data Species 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 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
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/35—Determination of transform parameters for the alignment of images, i.e. image registration using statistical methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/04—Indexing scheme for image data processing or generation, in general involving 3D image data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/32—Indexing scheme for image data processing or generation, in general involving image mosaicing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Computer Graphics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Ladders (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
This application provides a kind of point cloud registering optimization methods and device based on indoor three-dimensional data, this method comprises: obtaining the point cloud before the splicing that acquisition equipment acquires, and for pre-splicing outer ginseng matrix；Determine the point in described cloud there are ladder, and there are the ladder heights of the point of ladder；Based on pre-splicing outer ginseng matrix, the point that connection relationship will be present is divided into same connected domain, and there will be no the points of connection relationship to be divided into different connected domains；Using the ground comprising ground normal vector in the most connected domain of number and the smallest point of Y-axis angle as datum level；The ground of the point is overlapped with the datum level by adjusting the outer ginseng matrix of the point in the connected domain；The outer ginseng matrix of all the points in other connected domains is adjusted according to the ladder height of the point in the datum level and other connected domains；Splice described cloud using outer ginseng matrix adjusted.This method can be improved a cloud accuracy rate.
Description
Technical field
The present invention relates to field of computer technology, in particular to a kind of point cloud registering optimization side based on indoor three-dimensional data
Method and device.
Background technique
It needs to carry out data acquisition using acquisition equipment (camera etc.) in the indoor threedimensional model building process of the source of houses, obtain
Depth panorama sketch be that (using camera shooting point as origin, the position of each point in three dimensions is a large amount of discrete point datas
Discrete), and the specific environment of indoor shot determines and has a large amount of plane, for example, metope, desktop, ceiling and ground
Deng.
The problem of due to hardware device precision, in data acquisition, it is impossible to guarantee on collected same plane
Point can be also exactly in same plane in cloud, tentatively spliced by camera or existing some joining methods realized
Three-dimensional modeling, it is possible that the case where ground is crooked, out-of-flatness.
Summary of the invention
In view of this, the application provides a kind of point cloud registering optimization method and device based on indoor three-dimensional data, it can
Improve point cloud accuracy rate.
In order to solve the above technical problems, the technical solution of the application is achieved in that
A kind of point cloud registering optimization method based on indoor three-dimensional data is provided in one embodiment, this method comprises:
Point cloud before obtaining the splicing of acquisition equipment acquisition, and for pre-splicing outer ginseng matrix；
It determines in described cloud there are the point of ladder and described there are the ladder heights of the point of ladder；
It is used for pre-splicing outer ginseng matrix based on described, determines in described cloud and is closed between any two points with the presence or absence of connection
System；The point that connection relationship will be present is divided into same connected domain, and there will be no the points of connection relationship to be divided into different connections
In domain；
Using the ground comprising ground normal vector in the most connected domain of number and the smallest point of Y-axis angle as benchmark
Face；The ground of the point is overlapped with the datum level by adjusting the outer ginseng matrix of the point in the connected domain；
The outer of all the points in other connected domains is adjusted according to the ladder height of the point in the datum level and other connected domains
Join matrix；
Splice described cloud using outer ginseng matrix adjusted；
Wherein, the Y-axis is perpendicular to the ground, it is described there are the ground of the point of ladder be and acquisition equipment center position hang down
Straight difference in height is less than preset threshold, and with the smallest face of preset threshold range difference；It is the point height there is no the ground of ladder point
Minimum face.
In another embodiment, a kind of point cloud registering optimization device based on indoor three-dimensional data, the device packet are provided
It includes: acquiring unit, the first determination unit, the second determination unit, adjustment unit and concatenation unit；
The acquiring unit, the point cloud before splicing for obtaining acquisition equipment acquisition, and for pre-splicing outer ginseng
Matrix；
First determination unit, for determining in described cloud that the acquiring unit obtains there are the point of ladder, with
And described there are the ladder heights of the point of ladder；
Second determination unit is used for pre-splicing outer ginseng matrix for what is obtained based on the acquiring unit, determined
It whether there is connection relationship between any two points in point cloud；The point that connection relationship will be present is divided into same connected domain；It will not
There are the points of connection relationship to be divided into different connected domains；
The adjustment unit, for ground in the connected domains most comprising point number by second determination unit determination
The ground of normal vector and the smallest point of Y-axis angle is as datum level；Make this by adjusting the outer ginseng matrix of the point in the connected domain
The ground of point is overlapped with the datum level；In other connected domains determined according to the datum level and first determination unit
The ladder height of point adjust the outer ginseng matrixes of all the points in other connected domains；
The concatenation unit, for splicing described cloud using the adjustment unit outer ginseng matrix adjusted；Wherein, institute
State that Y-axis is perpendicular to the ground, it is described there are the ground of the point of ladder be and the vertical drop of acquisition equipment center position be less than it is pre-
If threshold value, and with the smallest face of preset threshold range difference；It is the minimum face of the point height there is no the ground of ladder point.
In another embodiment, a kind of computer readable storage medium is additionally provided, computer program is stored thereon with,
It is characterized in that, the method for the point cloud registering optimization method based on indoor three-dimensional data is realized when which is executed by processor
Step.
As can be seen from the above technical solution, there are the points of ladder in lookup point cloud in the application, and there are ladder points for calculating
Ladder height；And the point that connection will be present is divided into same connected domain, is found from comprising the most connected domain of number
Ground normal vector and the smallest point of Y-axis angle, using the ground that this puts as benchmark, to the outer ginseng matrixes of other points in cloud into
Row adjustment, re-starts splicing to a cloud using outer ginseng matrix adjusted, the point cloud registering for realizing indoor three-dimensional data is excellent
Change.This method can be improved the accuracy rate of splicing by adjusting outer ginseng matrix.And then overcome in three-dimensional modeling ground skew, no
Smooth problem.
Detailed description of the invention
Only illustratively description and explain the present invention for the following drawings, not delimit the scope of the invention:
Fig. 1 is the point cloud registering Optimizing Flow schematic diagram based on indoor three-dimensional data in the embodiment of the present application；
Fig. 2 is the schematic diagram of the corresponding ladder of point there are ladder height；
Fig. 3 is to determine the schematic diagram that whether there is connection relationship between point A and point B；
Fig. 4 a and Fig. 4 b are ground rotation and translation schematic diagram in the embodiment of the present invention；
Fig. 5 is the 3-D effect schematic diagram for acquiring equipment and splicing；
Fig. 6 is using method of adjustment provided by the present application to Fig. 5 3-D effect schematic diagram adjusted；
Fig. 7 is the apparatus structure schematic diagram for being applied to above-mentioned technology in the embodiment of the present application.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, the technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
Every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
The acquisition phase of three-dimensional data indoors will use special hardware device, acquire each shooting point in room
Data carry out each shooting point in acquisition equipment pre-splicing.
The outer ginseng matrix of point cloud and each point before being obtained with splicing in this way.
In the embodiment of the present application on the basis of pre-splicing contact cloud, outer ginseng matrix is adjusted, is realized to the point cloud spliced
The adjustment of plane.
For the point cloud to be spliced there may be multiple groups, such as the point cloud of some direction photographs, as one group of point cloud, this Shen
It please be similar to the processing mode of all groups of point clouds in embodiment by taking the processing mode for one group of point cloud as an example.
With reference to the accompanying drawing, it is described in detail and realizes the point cloud registering optimization based on indoor three-dimensional data in the embodiment of the present application
Process.
Realize that the device of the point cloud registering optimization based on indoor three-dimensional data can be for calculating in the embodiment of the present application
The equipment of processing capacity, such as a PC hereafter can referred to as optimize equipment for convenience.
Referring to Fig. 1, Fig. 1 is the point cloud registering Optimizing Flow schematic diagram based on indoor three-dimensional data in the embodiment of the present application.
Specific steps are as follows:
Step 101, optimization equipment obtains the point cloud before the splicing of acquisition equipment acquisition, and for pre-splicing outer ginseng square
Battle array.
The point cloud before splicing is obtained, and concrete mode the application for pre-splicing outer ginseng matrix is with no restrictions, such as
It can be and obtained from acquisition equipment by wired or wireless network, be also possible to obtain by way of disk copy.
Step 102, it determines in described cloud there are the point of ladder and described there are the ladder heights of the point of ladder.
Determine that there are the detailed processes of the point of ladder in point cloud in this step are as follows:
The first step, for each point in cloud, determine the minimum face of the point height；
Determine the minimum face of the point height, comprising:
Normal vector in plane corresponding to the point and Y-axis angle are less than the plane of predetermined angle threshold value as the first plane
Set；
Here plane corresponding to a point refers to all planes by the point.
It determines minimum as the point height with the acquisition maximum plane of equipment center vertical range in the first plane set
Face.
When obtaining the point cloud before splicing, the location information of present position when acquisition equipment acquires described cloud is obtained, that is, is adopted
Collect the coordinate information of the central point of equipment.
If the vertical height of the minimum face of second step, the point height and acquisition equipment center position is greater than preset threshold,
Determine that the point is the point there are ladder；
Third step is less than default threshold for the vertical drop for there are the point of ladder, determining and acquiring equipment center position
Value, and the ground with the smallest face of preset threshold range difference as the point；
For the point that ladder is not present, the minimum face of the height of the point is as ground.
4th step, the vertical height difference on the ground for calculating point and the minimum face of height, the ladder height as the point.
Referring to fig. 2, Fig. 2 is the schematic diagram of the corresponding ladder of point there are ladder height.With the lowest surfaces of a point in Fig. 2
With the difference in height of the center of acquisition equipment be greater than preset threshold (in Fig. 2 by taking 10cm as an example, can be according to adopting in concrete application
The preset threshold is arranged in collection equipment and the height on ground) for, then illustrate that there are ladder heights for the point, then finds the ground of the point
Face, using the difference in height of ground and lowest surfaces as the ladder height of the point.
Step 103, it is used for pre-splicing outer ginseng matrix based on described, determines in described cloud whether deposit between any two points
In connection relationship；The point that connection relationship will be present is divided into same connected domain；There will be no the points of connection relationship to be divided into not
In same connected domain.
Based on pre-splicing outer ginseng matrix in this step, determining in point cloud whether there is connection relationship between any two points,
Include:
By any two point in cloud: for first point (point A) and second point (point B)；
Point A is converted by the outer ginseng matrix of point A, point B is converted by the outer ginseng matrix of point B；
Determine whether intersect between transformed point A and point B, if so, determining that there are connection relationships between point A and point B；
Otherwise, it determines connection relationship is not present between point A and point B.
It is to determine the schematic diagram that whether there is connection relationship between point A and point B referring to Fig. 3, Fig. 3.Assuming that the midpoint Fig. 3 A
The content that outer ginseng matrix includes are as follows: rotate clockwise 90 °, 2 units of left.The content that the outer ginseng matrix of point B includes are as follows: inverse
Hour hands rotate 180 °, 1 unit of right translation.
By the point A after the outer ginseng matrixing of point A, and by the point B after the outer ginseng matrixing of point B, as seen from Figure 3
Two points after conversion are that there are connection relationships, that is to say, that the two points are intersections.
Point, which is divided into corresponding connected domain, to be accomplished in that
The point that connection relationship will be present establishes non-directed graph, and there are connection relationship, point 2 and points 3 there is connection and close for such as point 1 and 2
There are connection relationships for system, point 4 and point 5, and there are connection relationships for point 5 and point 6, then put 1, point 2 and point 3 establishes a non-directed graph, point
4, point 5 and point 6 establish a non-directed graph.
Each non-directed graph corresponds to a connected domain, and all the points in same non-directed graph are divided into a connected domain, no
It is divided into different connected domains with the point in non-directed graph.
Step 104, will make comprising the ground of ground normal vector in the most connected domain of number and the smallest point of Y-axis angle
For benchmark face；Outer ginseng matrix by adjusting other points in the connected domain is overlapped the ground of other points with the datum level.
The ground of point is defined as follows in the embodiment of the present application:
There are the ground of the point of ladder be with acquisition equipment center position vertical drop be less than preset threshold, and with it is pre-
If the smallest face of threshold distance difference；It is the minimum face of the point height there is no the ground of ladder point.
In the embodiment of the present application, the Y-axis is perpendicular to the ground；X0Z plane is parallel to the ground, and X, Y, Z are three-dimensional sit here
Mark, ground here can be ground corresponding to acquisition equipment.
According to above-mentioned treatment process in the embodiment of the present application, multiple connected domains can be established, are specifically located in the embodiment of the present application
When reason, can using comprising the most connected domain of point data as benchmark connected domain；
If the number comprising the most connected domain of number is greater than 1, one connected domain of random selection is connected to as benchmark
Domain；
Using the ground of ground normal vector in benchmark connected domain and the smallest point of Y-axis angle as datum level.
Outer ginseng matrix in this step by adjusting other points in the connected domain makes the ground of other points and the benchmark
Face is overlapped, comprising:
The ground normal vector and the smallest point of Y-axis angle are denoted as third point, except in the connected domain thirdly
Any point be denoted as the 4th point；
The outer ginseng matrix for adjusting the 4th point, is overlapped the ground of the 4th point and the ground of third point, specifically:
The minimum angle for determining the ground of the 4th point and the ground of third point, using the minimum angle degree more
Rotation angle in the outer ginseng matrix of new 4th point；
The ground rotation minimum angles of 4th point are parallel with the ground of third point, in specific implementation, if the
The ground of four points rotates clockwise parallel with the ground of third point for the positive direction of any axis, then updates outer join in matrix
Rotation angle when, the degree of minimum angle is added in former rotation angle；If the ground of the 4th point is for any axis
Positive rotation counterclockwise is parallel with the ground of third point, then the degree of minimum angle is reduced in former rotation angle.
Determine the height on the ground on the ground and third point using updated rotation postrotational 4th point of angle
Difference, and update using the difference in height translational movement in the outer ginseng matrix of the 4th point.
If the ground of the 4th point translates to Y-axis positive direction and is overlapped with the ground of third point, difference in height is negative value；
If the ground of the 4th point translates to Y-axis negative direction and is overlapped with the ground of third point, difference in height is positive value.
The adjustment of the outer ginseng matrix comprising each point in the most connected domain of point number is realized based on foregoing description.
Step 105, institute in other connected domains is adjusted according to the ladder height of the point in the datum level and other connected domains
Outer ginseng matrix a little.
Using the first connected domain as benchmark connected domain, for the ground of the third point in the first connected domain is benchmark face, come
It is adjusted for the outer ginseng matrix of all the points in the second connected domain, specific as follows:
Second connected domain is to include any connected domain except the most connected domain of a number.
By taking the outer ginseng adjustment of matrix of the 5th point and the 6th point in the second connected domain as an example, wherein the 5th point and
6th point is any two point in the second connected domain；5th point has confirmed that ladder height, the 6th point do not determine
For ladder point, ladder height is not present:
For the adjustment of the outer ginseng matrix of the 5th point are as follows:
The minimum angle for determining the ground of the 5th point and the ground of third point updates the 5th using the minimum angle
Rotation angle in the outer ginseng matrix of point；
The ground rotation minimum angles of 5th point are parallel with the ground of third point, in specific implementation, if the
The ground of five points rotates clockwise parallel with the ground of third point for the positive direction of any axis, then updates outer join in matrix
Rotation angle when, the degree of minimum angle is added in former rotation angle；If the ground of the 5th point is for any axis
Positive rotation counterclockwise is parallel with the ground of third point, then the degree of minimum angle is reduced in former rotation angle.
Determine the height on the ground on the ground and third point using updated rotation postrotational 5th point of angle
Difference, and the difference in height of use determination and the ladder height of the 5th point update the translational movement of the outer ginseng matrix of the 5th point.
In specific implementation, if using the ground of updated rotation postrotational 5th point of angle with Y-axis pros
It is overlapped to translation with the ground of third point, then the ground using updated rotation postrotational 5th point of angle and third
The difference in height on the ground of a point is negative value, when translating coincidence with Y-axis negative direction, the ground of the 5th point and the ground of third point
Difference in height be positive value.
For the adjustment of the outer ginseng matrix of the 6th point:
The minimum angle for determining the ground of the 6th point and the ground of third point updates the 6th using the minimum angle
Rotation angle in the outer ginseng matrix of point；
The ground rotation minimum angles of 6th point are parallel with the ground of third point, in specific implementation, if the
The ground of six points rotates clockwise parallel with the ground of third point for the positive direction of any axis, then updates outer join in matrix
Rotation angle when, the degree of minimum angle is added in former rotation angle；If the ground of the 6th point is for any axis
Positive rotation counterclockwise is parallel with the ground of third point, then the degree of minimum angle is reduced in former rotation angle.
If ladder height is not present in the 6th point, uses and be in same connected domain that there are the points of ladder with the 6th point
Ladder height of the ladder height as the 6th point, and more using the ladder height of determining difference in height and the 6th point
The translational movement of the outer ginseng matrix of new 6th point.
In specific implementation, if using the ground of updated rotation postrotational 6th point of angle with Y-axis pros
It is overlapped to translation with the ground of third point, then the ground using updated rotation postrotational 6th point of angle and third
The difference in height on the ground of a point is negative value, when translating coincidence with Y-axis negative direction, the ground of the 6th point and the ground of third point
Difference in height be positive value.
Step 106, splice described cloud using outer ginseng matrix adjusted.
Outer ginseng matrix in the embodiment of the present application by adjusting after splice it is pre-splicing before point cloud, realize for pre-splicing
Indoor three dimensional point cloud point cloud plane adjustment, and then realize indoor three-dimensional data point cloud registering optimization.
By adjusting the point cloud for having spliced completion is adjusted for the outer ginseng matrix of splicing in the embodiment of the present application, tune is used
Outer ginseng matrix after whole carries out splicing fusion to cloud, is generating triangle gridding, is completing the process of Model Reconstruction, can be according to existing
There is implementation realization.
Below with reference to specific example, detailed description realizes the process of point cloud plane adjustment:
Assuming that for one group of point cloud, there are five points: respectively point A, point B, point C, point D and point E.
Ladder height is determined whether there is for each point, it is assumed that determining point C, there are ladder heights, and ladder height is a,
Ladder height is not present in other points, then the lowest surfaces of point A, B, D and E are the ground of point A, B, D and E；The ground of point C is height
Difference be less than preset threshold, and with the smallest face of preset threshold range difference.
It determines and whether there is connection in above-mentioned five points between any two points, it is assumed that exist between point A, point B and point E and connect
It connects, is divided into a connected domain (the first connected domain), there is connection between point C and point D, be divided into a connected domain (second
Connected domain) in, connection relationship is not present between the point in the first connected domain midpoint and the second connected domain.
Since the first connected domain middle data is most, using the first connected domain as benchmark connected domain, it is assumed that the first connection
The ground normal vector and y-axis angle of domain midpoint A is minimum, then the ground of point of use A carries out the point in each connected domain as datum level
Outer ginseng matrix adjustment:
It is c ° that the ground level difference for calculating point A and point B, which is b cm rotation differential seat angle,；The ground level difference of point A and point E is
Tcm, and rotating differential seat angle is m °；
The ground level difference of point C and point A is k cm, and point C and point A differential seat angle are d °；The ground level difference of point D and point A is n
Cm, and rotating angle is z °.
For outer ginseng matrix (translation is y-axis direction)；
The original outer ginseng matrix of each point are as follows:
Point A: translation e cm rotates f °；
Point B: translation g cm rotates h °；
Point C: translation i cm rotates j °；
Point D: translation s cm rotates a °；
Point E: translation p cm rotates o °.
Each point outer ginseng matrix adjusted are as follows:
Point A: translation e cm rotates f °
Point B: translation g+b cm rotates (h+c) °
Point C: translation g+k+a cm rotates (j+d) °
Point D: translation s+k+n cm rotates (a+z) °；
Point E: translation p+t cm rotates (o+m) °.
Wherein, although ladder height is not present in point D, point D and point C belong to same connected domain, then the ladder height of point D
It is determined as identical as point C.
Fig. 4 a and Fig. 4 b are ground rotation and translation schematic diagram in the embodiment of the present invention.Wherein, Fig. 4 b be point B ground and
The ground point A angle schematic diagram；Fig. 4 a is the ground of point B and the difference in height schematic diagram on the ground point A.
Shown in foreknowledge Fig. 4 a, the ground of point B is ground B, and the ground of point A is ground A, and ground B rotates to flat with ground A
Row, the minimum angle for needing to rotate are assumed to be c °, using X-direction as axis shown in Fig. 4 a, rotate clockwise, then c ° of above-mentioned increase
When, increased angle value is positive value；If contrary, it is determined that be negative value.
Shown in goodbye Fig. 4 b, by taking ground A and postrotational ground B in Fig. 4 a as an example, postrotational ground B need to be along Y
The difference in height (absolute value of b) that axis positive direction translates two ground is overlapped with postrotational ground A, then outer to join the increased b of matrix
Value is negative value；If need to translate along Y-axis negative direction, the outer increased value of translational movement in matrix of joining is positive value.
By searching for the point in cloud there are ladder in the above embodiments of the present application, and there are the ladder of ladder point height for calculating
Degree；And the point that connection will be present is divided into same connected domain, finds ground normal direction from comprising the most connected domain of number
Amount is adjusted the outer ginseng matrix of other points in a cloud using the ground that this puts as benchmark with the smallest point of Y-axis angle,
Splicing is re-started to a cloud using outer ginseng matrix adjusted.This method can be improved splicing by adjusting outer ginseng matrix
Accuracy rate.And then overcome the problems, such as ground skew, out-of-flatness in three-dimensional modeling.
It is the 3-D effect schematic diagram for acquiring equipment and splicing referring to Fig. 5, Fig. 5.
As seen from Figure 5, subaerial part have many places be it is crooked, out-of-flatness, visual effect is bad.
Referring to Fig. 6, Fig. 6 is using method of adjustment provided by the present application to Fig. 5 3-D effect schematic diagram adjusted.
For Fig. 6 compared with Fig. 5, the marginal portion of entire 3 d effect graph is relatively flat, also compares alignment.
A kind of point cloud registering based on indoor three-dimensional data is also provided based on same inventive concept, in the embodiment of the present application
Optimize device.Referring to Fig. 7, Fig. 7 is the apparatus structure schematic diagram for being applied to above-mentioned technology in the embodiment of the present application.The device packet
It includes: acquiring unit 701, the first determination unit 702, the second determination unit 703, adjustment unit 704 and concatenation unit 705；
Acquiring unit 701, the point cloud before splicing for obtaining acquisition equipment acquisition, and for pre-splicing outer ginseng square
Battle array；
First determination unit 702 for determining in point cloud that acquiring unit 701 obtains there are the point of ladder, and exists
The ladder height of the point of ladder；
Second determination unit 703, the pre-splicing outer ginseng matrix for being obtained based on acquiring unit 701 are determined in point cloud
It whether there is connection relationship between any two points；The point that connection relationship will be present is divided into same connected domain；There will be no even
The point for connecing relationship is divided into different connected domains；
Adjustment unit 704, for ground method in the connected domains most comprising point number by the determination of the second determination unit 703
The ground of vector and the smallest point of Y-axis angle is as datum level；Make the point by adjusting the outer ginseng matrix of the point in the connected domain
Ground be overlapped with the datum level；The point in other connected domains determined according to the datum level and the first determination unit 702
Ladder height adjust the outer ginseng matrixes of all the points in other connected domains；
Concatenation unit 705, for using the outer ginseng matrix adjusted of adjustment unit 704 to splice described cloud；Wherein, Y-axis
It is perpendicular to the ground, it is the vertical drop with acquisition equipment center position less than preset threshold there are the ground of the point of ladder, and
With the smallest face of preset threshold range difference；It is the minimum face of the point height there is no the ground of ladder point.
Preferably,
First determination unit 702, specifically for there are the points of ladder in described cloud of determination, and there are the points of ladder
When ladder height, for each point in cloud, the minimum face of the point height is determined；If the minimum face of the point height is set with acquisition
The vertical height of standby center is greater than preset threshold, it is determined that the point is the point there are ladder；There are the points of ladder for calculating
The vertical height difference on ground and the minimum face of height, the ladder height as the point.
Preferably,
First determination unit 702, when specifically for the minimum face of the determining point height, by method in plane corresponding to the point
Vector and Y-axis angle are less than the plane of predetermined angle threshold value as the first plane set；Determine in the first plane set with acquisition
The maximum plane of the equipment center vertical range face minimum as the point height.
Preferably,
Second determination unit 703, specifically for based on pre-splicing outer ginseng matrix, determine in point cloud is between any two points
It is no there are when connection relationship, first point is converted by the outer ginseng matrix of first point, passes through the outer ginseng of second point
Matrix converts second point；Determine whether intersect between transformed first point and second point, if so, determining
There are connection relationships between first point and second point；Otherwise, it determines there is no connections between first point and second point
Relationship；Wherein, first point and second point are any two point in point cloud.
Preferably,
Adjustment unit 704 makes ground and the institute of the point specifically for the outer ginseng matrix by adjusting the point in the connected domain
When stating datum level coincidence, the ground normal vector and the smallest point of Y-axis angle are denoted as third point, by third in the connected domain
Any point except point is denoted as the 4th point；It determines the minimum angle on the ground of the 4th point and the ground of third point, uses
The degree of the minimum angle updates the rotation angle in the outer ginseng matrix of the 4th point；It determines and uses updated rotation angle
The difference in height on the ground on the ground and third point of postrotational 4th point, and the 4th point is updated using the difference in height
Translational movement in outer ginseng matrix.
Preferably,
Adjustment unit 704 is specifically used for adjusting it according to the ladder height of the point in the datum level and other connected domains
In its connected domain when the outer ginseng matrix of all the points, any in any connected domain except the most connected domain of a number will be included
Point is denoted as the 5th point；Determine the minimum angle on the ground of the 5th point and the ground of third point, more using the minimum angle
Rotation angle in the outer ginseng matrix of new 5th point；Determine the ground using updated rotation postrotational 5th point of angle
The difference in height in face and the ground of third point uses determining difference in height and the 5th if there are ladder heights for the 5th point
The ladder height of a point updates the translational movement of the outer ginseng matrix of the 5th point.
Preferably,
Adjustment unit 704, is further used for if it is determined that the 5th point is there is no ladder height, then use at the 5th point
Ladder height in same connected domain there are the ladder height of the point of ladder as the 5th point, and use determining height
The ladder height of difference and the 5th point updates the translational movement of the outer ginseng matrix of the 5th point.
The unit of above-described embodiment can integrate in one, can also be deployed separately；It can be merged into a unit, it can also
To be further split into multiple subelements.
In addition, the application also provides a kind of computer readable storage medium, it is stored thereon with computer program, the program quilt
The method and step of the point cloud registering optimization method based on indoor three-dimensional data as previously described is realized when processor executes.
In conclusion the application is by searching for the point in cloud there are ladder, and calculate that there are the ladder heights of ladder point；
And the point that connection will be present is divided into same connected domain, found from comprising the most connected domain of number ground normal vector with
The smallest point of Y-axis angle is adjusted the outer ginseng matrix of other points in a cloud using the ground that this puts as benchmark, uses
Outer ginseng matrix adjusted re-starts splicing to a cloud, realizes the point cloud registering optimization of indoor three-dimensional data.This method is logical
The outer ginseng matrix of adjustment is crossed, can be improved the accuracy rate of splicing.And then overcome the problems, such as ground skew, out-of-flatness in three-dimensional modeling.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Within mind and principle, any modification, equivalent substitution, improvement and etc. done be should be included within the scope of the present invention.
Claims (9)
1. a kind of point cloud registering optimization method based on indoor three-dimensional data, which is characterized in that this method comprises:
Point cloud before obtaining the splicing of acquisition equipment acquisition, and for pre-splicing outer ginseng matrix；
It determines in described cloud there are the point of ladder and described there are the ladder heights of the point of ladder；
It is used for pre-splicing outer ginseng matrix based on described, determines and whether there is connection relationship in described cloud between any two points；
The point that connection relationship will be present is divided into same connected domain, and there will be no the points of connection relationship to be divided into different connected domains
In；
Using the ground comprising ground normal vector in the most connected domain of number and the smallest point of Y-axis angle as datum level；It is logical
The outer ginseng matrix for crossing the point adjusted in the connected domain is overlapped the ground of the point with the datum level；
The outer ginseng square of all the points in other connected domains is adjusted according to the ladder height of the point in the datum level and other connected domains
Battle array；
Splice described cloud using outer ginseng matrix adjusted；
Wherein, the Y-axis is perpendicular to the ground, it is described there are the ground of the point of ladder be and acquisition equipment center position vertical height
Degree difference be less than preset threshold, and with the smallest face of preset threshold range difference；It is that the point height is minimum there is no the ground of ladder point
Face.
2. the method according to claim 1, wherein in described cloud of the determination, there are the points of ladder, and
It is described that there are the ladder heights of the point of ladder, comprising:
For each point in described cloud, the minimum face of the point height is determined；
If the vertical height of the minimum face of the point height and acquisition equipment center position is greater than preset threshold, it is determined that the point is to deposit
In the point of ladder；
Calculate described there are the vertical height difference on the ground of the point of ladder and the minimum face of height, the ladder as the point is high
Degree.
3. according to the method described in claim 2, it is characterized in that, the minimum face of the determination point height, comprising:
Normal vector in plane corresponding to the point and Y-axis angle are less than the plane of predetermined angle threshold value as the first planar set
It closes；
Determine face minimum as the point height with the acquisition maximum plane of equipment center vertical range in the first plane set.
4. the method according to claim 1, wherein described be used for pre-splicing outer ginseng matrix based on described, really
It pinpoints and whether there is connection relationship in cloud between any two points, comprising:
First point is converted by the outer ginseng matrix of first point, by the outer ginseng matrix of second point to second point
It is converted；
Determine whether intersect between transformed first point and second point, if so, determining first point and second point
Between there are connection relationships；Otherwise, it determines connection relationship is not present between first point and second point；
Wherein, first point and second point are any two point in point cloud.
5. method according to claim 1-4, which is characterized in that the point by adjusting in the connected domain
Outer ginseng matrix is overlapped the ground of the point with the datum level, comprising:
The ground normal vector and the smallest point of Y-axis angle are denoted as third point, by the connected domain thirdly except times
Any is denoted as the 4th point；
The minimum angle for determining the ground of the 4th point and the ground of third point updates the using the degree of the minimum angle
Rotation angle in the outer ginseng matrix of four points；
Determine the difference in height on the ground on the ground and third point using updated rotation postrotational 4th point of angle, and
The translational movement in the outer ginseng matrix of the 4th point is updated using the difference in height.
6. according to the method described in claim 5, it is characterized in that, described according in the datum level and other connected domains
The ladder height of point adjusts the outer ginseng matrix of all the points in other connected domains, comprising:
The 5th point will be denoted as comprising any point in any connected domain except the most connected domain of number；
The minimum angle for determining the ground of the 5th point and the ground of third point updates the 5th point using the minimum angle
Rotation angle in outer ginseng matrix；
Determine the difference in height on the ground on the ground and third point using updated rotation postrotational 5th point of angle, if
There are ladder heights for 5th point, then update the 5th point using the ladder height of determining difference in height and the 5th point
The translational movement of outer ginseng matrix.
7. according to the method described in claim 6, it is characterized in that, the method further includes:
If ladder height is not present in the 5th point, uses and be in same connected domain that there are the ranks of the point of ladder with the 5th point
Ladder height of the terraced height as the 5th point, and use the ladder height of determining difference in height and the 5th point update the
The translational movement of the outer ginseng matrix of five points.
8. a kind of point cloud registering based on indoor three-dimensional data optimizes device, which is characterized in that the device include: acquiring unit,
First determination unit, the second determination unit, adjustment unit and concatenation unit；
The acquiring unit, the point cloud before splicing for obtaining acquisition equipment acquisition, and for pre-splicing outer ginseng matrix；
First determination unit, for determining, there are the point of ladder, Yi Jisuo in described cloud that the acquiring unit obtains
State the ladder height of the point there are ladder；
Second determination unit is used for pre-splicing outer ginseng matrix for what is obtained based on the acquiring unit, determines a point cloud
It whether there is connection relationship between middle any two points；The point that connection relationship will be present is divided into same connected domain；It will be not present
The point of connection relationship is divided into different connected domains；
The adjustment unit, for ground normal direction in the connected domains most comprising point number by second determination unit determination
The ground of amount and the smallest point of Y-axis angle is as datum level；Make the point by adjusting the outer ginseng matrix of the point in the connected domain
Ground is overlapped with the datum level；The point in other connected domains determined according to the datum level and first determination unit
Ladder height adjust the outer ginseng matrixes of all the points in other connected domains；
The concatenation unit, for splicing described cloud using the adjustment unit outer ginseng matrix adjusted；Wherein, the Y
Axis is perpendicular to the ground, described there are the ground of the point of ladder to be and the vertical drop of acquisition equipment center position is less than default threshold
Value, and with the smallest face of preset threshold range difference；It is the minimum face of the point height there is no the ground of ladder point.
9. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is held by processor
Claim 1-7 described in any item method and steps are realized when row.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910270158.3A CN110111378B (en) | 2019-04-04 | 2019-04-04 | Point cloud registration optimization method and device based on indoor three-dimensional data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910270158.3A CN110111378B (en) | 2019-04-04 | 2019-04-04 | Point cloud registration optimization method and device based on indoor three-dimensional data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110111378A true CN110111378A (en) | 2019-08-09 |
| CN110111378B CN110111378B (en) | 2021-07-02 |
Family
ID=67485152
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910270158.3A Active CN110111378B (en) | 2019-04-04 | 2019-04-04 | Point cloud registration optimization method and device based on indoor three-dimensional data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110111378B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110866449A (en) * | 2019-10-21 | 2020-03-06 | 北京京东尚科信息技术有限公司 | Method and device for identifying target object in road |
| CN111310818A (en) * | 2020-02-10 | 2020-06-19 | 贝壳技术有限公司 | Feature descriptor determining method and device and computer-readable storage medium |
| CN111325779A (en) * | 2020-02-07 | 2020-06-23 | 贝壳技术有限公司 | Point cloud registration method and device, electronic equipment and storage medium |
| CN115479545A (en) * | 2022-10-28 | 2022-12-16 | 博众精工科技股份有限公司 | Height difference measuring method and device, terminal and readable storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103322911A (en) * | 2013-05-27 | 2013-09-25 | 吴晓军 | Characteristic building elevation surveying and mapping method based on images |
| US20150156415A1 (en) * | 2011-12-30 | 2015-06-04 | Google Inc. | Multiplane Panoramas of Long Scenes |
| JP6091560B2 (en) * | 2014-07-25 | 2017-03-08 | 株式会社東芝 | Image analysis method |
| CN108981604A (en) * | 2018-07-11 | 2018-12-11 | 天津工业大学 | A kind of precision component three-dimensional overall picture measurement method based on line laser |
| CN109035422A (en) * | 2018-06-06 | 2018-12-18 | 链家网（北京）科技有限公司 | In a kind of extraction chamber in threedimensional model plane domain method and system |
-
2019
- 2019-04-04 CN CN201910270158.3A patent/CN110111378B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150156415A1 (en) * | 2011-12-30 | 2015-06-04 | Google Inc. | Multiplane Panoramas of Long Scenes |
| CN103322911A (en) * | 2013-05-27 | 2013-09-25 | 吴晓军 | Characteristic building elevation surveying and mapping method based on images |
| JP6091560B2 (en) * | 2014-07-25 | 2017-03-08 | 株式会社東芝 | Image analysis method |
| CN109035422A (en) * | 2018-06-06 | 2018-12-18 | 链家网（北京）科技有限公司 | In a kind of extraction chamber in threedimensional model plane domain method and system |
| CN108981604A (en) * | 2018-07-11 | 2018-12-11 | 天津工业大学 | A kind of precision component three-dimensional overall picture measurement method based on line laser |
Non-Patent Citations (2)
| Title |
|---|
| LACHHANI, K, DUAN, J, BAGHSIAHI: "Error Metric for Indoor 3D Point Cloud Registration", 《IRISH MACHINE VISION & IMAGE PROCESSING CONFERENCE》 * |
| 赵凯，徐友春，王任栋: "一种城市环境三维点云配准的", 《光电工程》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110866449A (en) * | 2019-10-21 | 2020-03-06 | 北京京东尚科信息技术有限公司 | Method and device for identifying target object in road |
| CN111325779A (en) * | 2020-02-07 | 2020-06-23 | 贝壳技术有限公司 | Point cloud registration method and device, electronic equipment and storage medium |
| CN111310818A (en) * | 2020-02-10 | 2020-06-19 | 贝壳技术有限公司 | Feature descriptor determining method and device and computer-readable storage medium |
| CN115479545A (en) * | 2022-10-28 | 2022-12-16 | 博众精工科技股份有限公司 | Height difference measuring method and device, terminal and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110111378B (en) | 2021-07-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110111378A (en) | A kind of point cloud registering optimization method and device based on indoor three-dimensional data | |
| US11200734B2 (en) | Method for reconstructing three-dimensional space scene based on photographing | |
| US20220028156A1 (en) | Generating Floor Maps For Buildings From Automated Analysis Of Visual Data Of The Buildings' Interiors | |
| CN1669069B (en) | System for texturizing electronic representations of objects | |
| CN108230247B (en) | Generation method, device, equipment and the computer-readable storage medium of three-dimensional map based on cloud | |
| CN103279983B (en) | The modeling method of China Tang dynasty style ancient building | |
| CN110163968A (en) | RGBD camera large-scale three dimensional scenario building method and system | |
| Pan et al. | Real-time collision detection and distance computation on point cloud sensor data | |
| CN102263933A (en) | Intelligent monitoring method and device | |
| JP6353175B1 (en) | Automatically combine images using visual features | |
| KR20110133587A (en) | A method for determining the relative position of a first and a second imaging device and devices therefore | |
| US9507749B2 (en) | Communication system and method of generating geographic social networks in virtual space | |
| CN111798562B (en) | Virtual building space building and roaming method | |
| Xu et al. | 3D Reconstruction system for collaborative scanning based on multiple RGB-D cameras | |
| CN113393577B (en) | Oblique photography terrain reconstruction method | |
| Chen et al. | Maximal matching of 3-D points for multiple-object motion estimation | |
| Faust et al. | Real-time global data model for the digital earth | |
| WO2021124286A1 (en) | Location determination and mapping with 3d line junctions | |
| US20230368482A1 (en) | Registration of 3d augmented scene to structural floor plans | |
| CN106296650A (en) | A kind of laser point cloud method for registering and device | |
| Araújo et al. | A Descriptive Geometry Construction of VR panoramas in Cubical Spherical Perspective | |
| CN204028993U (en) | A kind of object digital three-dimensional reconstruction apparatus scanned based on light Gate | |
| CN112948518A (en) | Object processing method and device, electronic equipment and computer storage medium | |
| Araújo et al. | How to draw a virtual cubical perspective box | |
| Zhang et al. | Research on Influence of Photogrammetry in Digital High Precision Modeling Technology on Production Efficiency of Ceramic Enterprises |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |