CN109448106A - Fruit point cloud method for reconstructing based on spherical coordinates - Google Patents
Fruit point cloud method for reconstructing based on spherical coordinates Download PDFInfo
- Publication number
- CN109448106A CN109448106A CN201811207363.7A CN201811207363A CN109448106A CN 109448106 A CN109448106 A CN 109448106A CN 201811207363 A CN201811207363 A CN 201811207363A CN 109448106 A CN109448106 A CN 109448106A
- Authority
- CN
- China
- Prior art keywords
- longitude
- latitude
- central axis
- point cloud
- fruit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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 transformation in the plane of the image
- G06T3/60—Rotation of a whole image or part thereof
- G06T3/604—Rotation of a whole image or part thereof using a CORDIC [COordinate Rotation Digital Compute] device
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Medicines Containing Plant Substances (AREA)
- Image Processing (AREA)
Abstract
The present invention provides a kind of fruit point cloud method for reconstructing based on spherical coordinates, includes the following steps: step 1, and based on three-dimensional fruit original point cloud data, a cloud rectangular co-ordinate is switched to spherical coordinates；Step 2, denoising is carried out to point cloud data；Step 3, the longitude and latitude of oblate spheroid central axis and the longitude and latitude of prolate sphere central axis are calculated to feature when being parallel to frontal plane of projection based on oblate spheroid central axis and prolate sphere center axis rotation；Step 4, the longitude and latitude of longitude and latitude and prolate sphere central axis based on oblate spheroid central axis, rebuilds fruit using B-spline surface.Computational efficiency of the present invention is high, and denoising effect is good, can be applied to the spheroidal fruit of class and rebuilds.
Description
Technical field
The invention belongs to object reconstruction technical fields, and in particular to a kind of fruit point cloud reconstruction side based on spherical coordinates
Method.
Background technique
Object reconstruction method based on three-dimensional point cloud has the characteristics that reconstruction precision is high, but the spatial relationship of point cloud data
Computational processing is big, low efficiency.Currently, the formative method of the fruit based on cloud mostly uses general points cloud processing side
Formula is handled for the characteristics of fruit shapes.
Summary of the invention
The object of the present invention is to provide a kind of fruit point cloud method for reconstructing based on spherical coordinates, for fruit point
Cloud feature, can faster and accurately degeneracy original point cloud, efficiently rebuild fruit.
The embodiment provides a kind of fruit point cloud method for reconstructing based on spherical coordinates, including walk as follows
It is rapid:
Step 1, based on three-dimensional fruit original point cloud data, a cloud rectangular co-ordinate is switched into spherical coordinates；
Step 2, denoising is carried out to point cloud data；
Step 3, it is calculated based on oblate spheroid central axis and prolate sphere center axis rotation to feature when being parallel to frontal plane of projection
The longitude and latitude of oblate spheroid central axis and the longitude and latitude of prolate sphere central axis；
Step 4, the longitude and latitude of longitude and latitude and prolate sphere central axis based on oblate spheroid central axis, using B-spline surface pair
Fruit is rebuild.
Further, step 1 includes:
In the point cloud data deposit single linked list that will acquire；
The center point coordinate for calculating point cloud recalculates a cloud coordinate after point cloud center translation to rectangular co-ordinate origin
It is stored in single linked list；
A cloud rectangular co-ordinate is switched into spherical coordinates value.
Further, step 2 includes:
Noise is filtered using the median filtering of sort method filter method.
Further, step 3 includes:
It calculates the central point through dotted line and surrounds circle；
Calculate warp and circular difference degree c；
For a fruit, if latitude differs two of 180 ° through minimum widith w between dotted lineminCorresponding c is cmin, between warp
Maximum width wmaxCorresponding c is cmax；
Work as cmin>cmaxWhen, which is oblate spheroid, wminCorresponding latitude is the latitude φ of center axis；
Work as cmin<cmaxWhen, which is prolate sphere, wmaxCorresponding latitude is the latitude φ of center axis；
Work as wmin=wmaxWhen, central axis is in vertical direction, warp height h > wminWhen, which is prolate sphere；h<wminWhen,
For oblate spheroid；H=wminWhen be spherosome.
Further, step 3 further include:
It will be rotated to through line width most in the plane around central point through corresponding two when line width maximum through dotted line
The longitude of hour is determined as the initial longitude θ of prolate sphere central axis0；
By initial longitude θ0It is set as 0 °, recalculates the longitude of the longitude points of latitude b=φ, φ+180；
On the same latitude, 5 adjacent longitude points are carried out smooth；
Encirclement circle of the prolate sphere through dotted line is calculated, round difference degree c is calculated, each non-empty longitude points is calculated and surrounds and justify
Distance dab；
D is obtained in+180 latitude of b=φ and b=φ respectivelyabThe longitude points of > c are found and straight line with least square method
The G-bar k of the smallest two fitting a straight lines of horizontal distance；
△ θ, longitude θ=θ of prolate sphere central axis are calculated based on slope k0+△θ。
Further, step 3 further include:
It will be rotated to through line width most in the plane around center dotted line through corresponding two when line width minimum through putting
Longitude when big is determined as the initial longitude θ of oblate spheroid central axis0；
By initial longitude θ0It is set as 0 °, recalculates the longitude of the longitude points of latitude b=φ, φ+180；
On the same latitude, 5 adjacent longitude points are carried out smooth；
Encirclement circle of the oblate spheroid through dotted line is calculated, round difference degree c is calculated, each non-empty longitude points is calculated and surrounds and justify
Distance dab；
Respectively in+180 latitude of b=φ and b=φ and within the scope of 0 °≤a≤45 ° of longitude and 135 °≤a≤180 °, obtain
Take dab> c non-empty longitude points are rotated by 90 ° with the smallest two fitting a straight lines of vertical range of least square method searching and straight line
G-bar k；
△ θ, longitude θ=θ of oblate spheroid central axis are calculated based on slope k0+△θ。
Further, step 4 includes:
Cloud rotation transformation and a denoising are carried out to original point cloud data；
Null point is filled according to the coordinate of non-null point within the scope of null point neighborhood longitude and latitude；
It is further denoised using mean filter；
Fruit reconstruction is carried out using 2 B-spline surfaces of uniform period；
The boundary of each fettucelle of B-spline surface is handled, B-spline surface is closed.
Compared with prior art the beneficial effects of the present invention are:
Computational efficiency of the present invention is high, and denoising effect is good, can be applied to the spheroidal fruit of class and rebuilds.
Detailed description of the invention
Fig. 1 is the orthographic drawing that the present invention uses three spheroidal original fruit point cloud datas；
Fig. 2 is the point cloud chart that three fruit latitude b are 0 degree and 180 degree；
Fig. 3 is the adjacency list of R [a] [b]；
Fig. 4 is the filter result adjacency list of R [a] [b]；
Fig. 5 is the point cloud in Fig. 1 after fruit denoising；
Fig. 6 is after Fig. 2 degeneracy through point and line chart；
Fig. 7 is the shape of sphere different rotary angle；
Fig. 8 is different fruit different latitudes through the postrotational orthographic drawing of dotted line；
Fig. 9 is three fruits through dotted line maximum and minimum widith and latitude；
Figure 10 is through dotted line smallest enclosing circle and round difference degree；
Figure 11 is that oblate spheroid is rotated through dotted line clockwise around central point；
Figure 12 is that prolate sphere is rotated through dotted line clockwise around central point；
Figure 13 is the initial longitude of fruit central axis；
Figure 14 is prolate sphere central axis longitude calculating process；
Figure 15 is oblate spheroid central axis longitude calculating process；
Figure 16 is fruit point cloud of the central axis in vertical direction；
Figure 17 is denoising consequence real point cloud orthographic drawing；
Figure 18 is the point cloud orthographic drawing that weft is 0 ° and 180 °；
Figure 19 is that weft is point cloud orthographic drawing after 0 ° and 180 ° of filling points；
Figure 20 is that weft is point cloud orthographic drawing after 0 ° and 180 ° of degeneracys；
Figure 21 is that weft is point cloud orthographic drawing after 0 ° and 180 ° filtering；
Figure 22 is the grid chart of fruit；
Figure 23 is B-spline fettucelle；
Figure 24 is not closed apple；
Figure 25 is the reconstruction figure of three fruit different angles.
Specific embodiment
The present invention is described in detail for each embodiment shown in reference to the accompanying drawing, but it should be stated that, these
Embodiment is not limitation of the present invention, those of ordinary skill in the art according to these embodiments made by function, method,
Or equivalent transformation or substitution in structure, all belong to the scope of protection of the present invention within.
A kind of fruit point cloud method for reconstructing based on spherical coordinates is present embodiments provided, is included the following steps:
Step 1, based on three-dimensional fruit original point cloud data, a cloud rectangular co-ordinate is switched into spherical coordinates；
Step 2, denoising is carried out to point cloud data；
Step 3, it is calculated based on oblate spheroid central axis and prolate sphere center axis rotation to feature when being parallel to frontal plane of projection
The longitude and latitude of oblate spheroid central axis and the longitude and latitude of prolate sphere central axis；
Step 4, the longitude and latitude of longitude and latitude and prolate sphere central axis based on oblate spheroid central axis, using B-spline surface pair
Fruit is rebuild.
The fruit point cloud method for reconstructing computational efficiency based on spherical coordinates is high, and denoising effect is good, can be applied to class
Spheroidal fruit is rebuild.
In the present embodiment, step 1 includes:
In the point cloud data deposit single linked list that will acquire；
The center point coordinate for calculating point cloud recalculates a cloud coordinate after point cloud center translation to rectangular co-ordinate origin
It is stored in single linked list；
A cloud rectangular co-ordinate is switched into spherical coordinates value.
In the present embodiment, step 2 includes:
Noise is filtered using the median filtering of sort method filter method.
In the present embodiment, step 3 includes:
It calculates the central point through dotted line and surrounds circle；
Calculate warp and circular difference degree c；
For a fruit, if latitude differs two of 180 ° through minimum widith w between dotted lineminCorresponding c is cmin, between warp
Maximum width wmaxCorresponding c is cmax；
Work as cmin>cmaxWhen, which is oblate spheroid, wminCorresponding latitude is the latitude φ of center axis；
Work as cmin<cmaxWhen, which is prolate sphere, wmaxCorresponding latitude is the latitude φ of center axis；
Work as wmin=wmaxWhen, central axis is in vertical direction, warp height h > wminWhen, which is prolate sphere h < wminWhen,
For oblate spheroid；H=wminWhen be spherosome.
In the present embodiment, step 3 further include:
It will be rotated to through line width most in the plane around center dotted line through corresponding two when line width maximum through putting
The longitude of hour is determined as the initial longitude θ of prolate sphere central axis0；
By initial longitude θ0It is set as 0 °, recalculates the longitude of the longitude points of latitude b=φ, φ+180；
On the same latitude, 5 adjacent longitude points are carried out smooth；
Encirclement circle of the prolate sphere through dotted line is calculated, round difference degree c is calculated, each non-empty longitude points is calculated and surrounds and justify
Distance dab；
D is obtained in+180 latitude of b=φ and b=φ respectivelyabThe longitude points of > c are found and straight line with least square method
The G-bar k of the smallest two fitting a straight lines of horizontal distance；
△ θ, longitude θ=θ of prolate sphere central axis are calculated based on slope k0+△θ。
In the present embodiment, step 3 further include:
It will be rotated to through line width most in the plane around central point through corresponding two when line width minimum through dotted line
Longitude when big is determined as the initial longitude θ of oblate spheroid central axis0；
By initial longitude θ0It is set as 0 °, recalculates the longitude of the longitude points of latitude b=φ, φ+180；
On the same latitude, 5 adjacent longitude points are carried out smooth；
Encirclement circle of the oblate spheroid through dotted line is calculated, round difference degree c is calculated, each non-empty longitude points is calculated and surrounds and justify
Distance dab；
Respectively in+180 latitude of b=φ and b=φ and within the scope of 0 °≤a≤45 ° of longitude and 135 °≤a≤180 °, obtain
Take dab> c non-empty longitude points are rotated by 90 ° with the smallest two fitting a straight lines of vertical range of least square method searching and straight line
G-bar k；
△ θ, longitude θ=θ of oblate spheroid central axis are calculated based on slope k0+△θ。
In the present embodiment, step 4 includes:
Cloud rotation transformation and a denoising are carried out to original point cloud data；
Null point is filled according to the coordinate of non-null point within the scope of null point neighborhood longitude and latitude；
It is further denoised using mean filter；
Fruit reconstruction is carried out using 2 B-spline surfaces of uniform period；
The boundary of each fettucelle of B-spline surface is handled, B-spline surface is closed.
Invention is further described in detail below.
1, rectangular co-ordinate switchs to spherical coordinates
If three-dimensional fruit point cloud data is P (xi,yi,zi) (i=0,1,2 ..., n-1).N is data amount check.Use the right hand
Coordinate system, horizontal direction is X-axis to the right, and direction is Y-axis downward vertically.Programming environment of the invention is Visual C++
6.0。
(1) point cloud data is read
Since point cloud data amount is larger, be not suitable for being saved in memory with array, Yi Caiyong chain sheet form.According to 3-D scanning
Point cloud data is stored in single linked list P by the point cloud data text file that instrument obtains, and the structural body type of each point is as follows:
(2) point cloud data is shown
By fruit point cloud orthographic projection to xoy coordinate plane:
Fig. 1 is the orthographic drawing (data unit that the present invention uses three spheroidal original fruit point cloud datas
For a pixel), wherein (a) is Nanfeng orange, 5733797 points；It (b) is bergamot pear, 4574689 points；It (c) is bergamot pear,
1000001 points.
(3) center point coordinate is calculated
The central point of point cloud calculates as follows:
After cloud center translation to rectangular co-ordinate origin being put, recalculate in a cloud coordinate deposit single linked list P.
(4) rectangular co-ordinate value switchs to spherical coordinates value
The conversion formula that cloud rectangular co-ordinate is switched to spherical coordinates is as follows:
Wherein: longitude 0≤a≤180 °, latitude 0≤b≤360 °, the increment of a and b takes 1 ° here.The rectangular co-ordinate of point cloud
After switching to spherical coordinates, be equivalent to and unordered rectangular co-ordinate point cloud switched into orderly gridding point, can quickly find out a little with
Relationship between point, facilitates subsequent processing.
2, point cloud data denoises
(1) storage mode of spherical coordinates value
Due to the point cloud data comparatively dense of acquisition, after rectangular co-ordinate (x, y, z) is converted to spherical coordinates (a, b, r), have more
Mapping of a different radial radius r to (a, a b) value, for fruit point cloud, (a, a b) value only needs a corresponding r
Value, the corresponding point of extra r can be used as noise and remove, and Fig. 2 is that three fruit latitude b are 0 ° and 180 ° through dotted line.
It can be seen that there is the data point of different numbers on each longitude and latitude, stored using adjacency list, node class
Type are as follows:
Each theodolite place exists in two-dimentional array of pointers:
struct pr*R[181][361]；
When corresponding to same (a, the b) value for different radial radius r, using single linked list, the rectangular co-ordinate of cloud will put
When switching to spherical coordinates, and it is sequentially inserted into single linked list from small to large by radius r value.As shown in Figure 3.
During rectangular co-ordinate switchs to spherical coordinates, for the longitude and latitude point (namely null point) being not transitioning to, at it
The node is set to NULL here by being marked at node, if the longitude and latitude of null point is respectively i and j, then R [i] [j]=
NULL。
(2) denoising
From figure 2 it can be seen that nearby had multiple noises through dotted line, some noises from through dotted line there are also with a distance from some, root
According to probability statistics rule, the point fallen near fruit surface is centainly more than noise, therefore, utilizes the intermediate value of sort method filter method
Filtering, can substantially filter out noise.
Median filtering is carried out to corresponding adjacent chained list all in R [181] [361], for example, to r1,r1,…,rnCarry out intermediate value filter
Wave:
Due to ri(i=1,2 ... n) have sorted, by rn/2In corresponding nodal value deposit R [a] [b] (Fig. 4).
Fig. 5 is the point cloud in Fig. 1 after fruit denoising.It can clearly be seen that the noise of periphery substantially eliminates.In this example
In, the point number after each fruit denoising is up to 180*360=64800 point, can adjust longitude and latitude according to the size of fruit
The interval of degree when fruit is smaller, can increase the interval of longitude and latitude, when fruit is larger, can reduce the interval of longitude and latitude.
Fig. 6 is after Fig. 2 is denoised through point and line chart.As can be seen from the figure noise spot is substantially eliminated after denoising.Although
Point cloud data is more, but setting interval longitude and latitude at, do not ensure that centainly have put accordingly it is corresponding, so can exist compared with
A small amount of null point, but the calculating of central axis is not influenced.
The result of point cloud data denoising are as follows: (xab, yab, zab) (a=0,1 ... 180, b=0,1 ... 360).
3, point cloud data central axis calculates
(1) feature of central axis
For the fruit of approximate prolate sphere and oblate spheroid, no matter how initial position is put, by different latitude through point
Line can obtain an angle direction of central shaft, as shown in Figure 7.
For oblate spheroid, when central axis (black thick vertical line) is in vertical direction, after being converted to spherical coordinate, oblate spheroid around
When vertical direction rotates, the shape through dotted line of two latitudes of outermost is identical after orthographic projection, such as the dotted line of Fig. 7 (a)；In the middle
Mandrel in the horizontal direction inside-out (or outside in) when, after being converted to spherical coordinate, oblate spheroid is (vertical empty around vertical direction
Line, similarly hereinafter) when rotation, after orthographic projection the width through dotted line of two latitudes of outermost can change (Fig. 7 (b), (c),
(d)), when central axis goes to horizontal direction, two orthographic projection width through dotted line are minimum (Fig. 7 (d))；When central axis is in rectangle
Xiang Shi, after being converted to spherical coordinate, when oblate spheroid is rotated around vertical direction, outermost two latitudes through dotted line after orthographic projection
Width can also change (Fig. 7 (e), (f), (g)), when central axis, which is gone to, is parallel to frontal plane of projection, two through dotted line just
Projection width is also minimum (Fig. 7 (g)).Therefore it can be concluded that, when the center axis rotation of oblate spheroid is to when being parallel to frontal plane of projection, just
The width through dotted line of two latitudes of its outermost is minimum after projection.
For prolate sphere, when central axis (black thick vertical line) is in vertical direction, after being converted to spherical coordinate, prolate sphere around
When vertical direction rotates, the shape through dotted line of two latitudes of outermost is identical after orthographic projection, such as the dotted line of Fig. 7 (h)；In the middle
Mandrel is in tilted direction, after being converted to spherical coordinate, when prolate sphere is rotated around vertical direction, and two latitudes of outermost after orthographic projection
The width through dotted line can change (Fig. 7 (i), (j), (k)), when central axis, which is gone to, is parallel to frontal plane of projection, two warp
The orthographic projection width of dotted line is maximum (Fig. 7 (k)).When central axis in the horizontal direction inside-out (or outside in) when, be converted to
After spherical coordinate, when prolate sphere is rotated around vertical direction, the width through dotted line of two latitudes of outermost can also be sent out after orthographic projection
Changing (Fig. 7 (l), (m), (n)), when central axis, which is gone to, is parallel to frontal plane of projection, two orthographic projection width through dotted line
Maximum (Fig. 7 (n))；It can be concluded that when the center axis rotation of prolate sphere is to when being parallel to frontal plane of projection, its outermost after orthographic projection
The width through dotted line of two latitudes is maximum.
The latitude feature of oblate spheroid and the central axis of prolate sphere according to other feature on the contrary, need to be distinguished.It is either flat
Sphere or prolate sphere, when central axis, which rotates to, is parallel to frontal plane of projection, two round degree through dotted line compare other rotations
Gyration (latitude) is the smallest.It is rotated after the latitude of central axis has been determined, then along longitudinal, until two through dotted line
Width maximum (oblate spheroid) is minimum (prolate sphere).
(2) latitude of central axis
For three fruits in Fig. 5, two latitudes of 180 ° of difference is taken to rotate θ angle around Y-axis through dotted line:
Fig. 8 (a), (b), (c) be respectively two latitudes of different fruits through dotted line around Y-axis rotate the angle θ (θ=0,30 ...,
150) orthographic drawing, wherein (a) is Nanfeng orange, (b) is apple, (c) is bergamot pear.The number of lower section is the width through dotted line
It spends w (as unit of pixel, similarly hereinafter).
From Fig. 8 (a), (b) as can be seen that according to through a line width minimum (wmin=58 and wmin=85) latitude, can
To determine the latitude similar to the Nanfeng orange of oblate spheroid and the central axis of apple.From Fig. 8 (c) as can be seen that according to through dotted line
Width maximum (wmax=88) latitude can determine the latitude of similar prolate sphere central axis.Each rotation angle is set as 3 °,
It can obtain minimum widith w between the warp of three fruitsminWith maximum width wmaxThrough dotted line and corresponding latitude φ, such as Fig. 9 institute
Show.
(3) oblate spheroid and prolate sphere are distinguished
For a fruit, the central point (x through dotted line in Fig. 9 is calculated first0,y0):
In formula: m is b=φ and b=φ+180 through null point coordinate (x ' non-on dotted lineab,y’ab) number.
Then the encirclement circle in Fig. 9 through dotted line is calculated, radius is through null point coordinate points non-on dotted line and central point (x0,
y0) maximum distance:
Such as Figure 10, the circle of warp periphery is to surround circle.
Finally calculate warp and circular difference degree
Number below Figure 10 middle longitude is the round difference degree c through dotted line.C=0 expression be it is round, c is smaller, explanation
Closer to circle.
Comparison diagram 9 and Figure 10, for oblate spheroid tangerine orange and apple, minimum its round difference degree through line width is maximum,
For prolate sphere bergamot pear, maximum its round difference degree through line width is maximum.
For a fruit, if wminCorresponding c is cmin,wmaxCorresponding c is cmax。
Work as cmin>cmaxWhen, it is oblate spheroid, wminCorresponding latitude is the latitude φ of center axis.
Work as cmin<cmaxWhen, it is prolate sphere, wmaxCorresponding latitude is the latitude φ of center axis.
Work as wmin=wmaxWhen, central axis is in vertical direction, through dotted line height h > wminWhen, it is prolate sphere；h<wminWhen, it is flat
Sphere；H=wminWhen be spherosome.
(4) longitude of central axis
1) the initial longitude of central axis
For oblate spheroid, by wminIt is corresponding through dotted line in the plane clockwise around central point (x0,y0) rotate by a certain angle
α:
X "=(xab’-x0)cosα-(yab’-y0)sinα+x0
Y "=(xab’-x0)sinα+(yab’-y0)cosα+y0(6)
When through line width maximum, corresponding longitude is exactly the initial longitude of central axis.Such as Figure 11, for oblate spheroid,
It is rotated by 360 ° at regular intervals, selects longitude θ when width maximum0。
Such as Figure 12, for prolate sphere, by wmaxCorresponding two through dotted line between the plane clockwise around central is pressed centainly
Every being rotated by 360 °, longitude when width minimum is selected, is exactly the initial longitude θ of central axis0。
Figure 13 is the initial longitude by the central axis of 3 ° of interval acquiring.Because of noise spot and fruit shapes not exclusively rule
It influences, needs to correct the initial longitude of central axis.
2) longitude through dotted line is recalculated
Formula (1) is simplified as, for Figure 13, recalculate the longitude points of two latitudes longitude a (b=φ,
φ+180)。
Corresponding points coordinate are as follows: (X 'ab, Y 'ab, Z 'ab) (a=0,1 ... 180, b=φ, φ+180).
3) smoothly through dotted line
On the same latitude, 5 adjacent longitude points are carried out smoothly, such as following formula.
In formula: n is through null point coordinate (X ' non-on dotted lineab,Y’ab) number.
4) prolate sphere central axis longitude
Encirclement circle (Figure 14 (a)) of the prolate sphere through dotted line is calculated using formula (3) formula (4), is calculated and is justified using formula (5)
Shape difference degree c, then calculate each non-empty longitude points and surround round distance dab:
Obtain dabThe longitude points (Figure 14 (b), generally in the right and left) of > c respectively carry out b=φ and b=φ+180 straight
Line fitting is found with the horizontal distance of straight line most since the actual direction of prolate sphere has been approximate direction with least square method
The slope of small fitting a straight line:
Two fitting a straight lines are average as shown in Figure 14 (c).This direction is exactly the direction of central axis.K is used using formula (7)
△ θ is calculated instead of y/x:
Longitude θ=θ of central axis0+△θ.Center axis rotation is to vertical direction such as Figure 14 (d).Final bergamot pear central axis
Longitude and latitude is (85 °, 78 °).
5) oblate spheroid central axis longitude
Encirclement circle (Figure 15 (a)) of the oblate spheroid through dotted line similarly is calculated using formula (3) formula (4), is counted using formula (5)
Round difference degree c is calculated, then calculates each non-empty longitude points and surrounds round distance dab:
For oblate spheroid, dabThe longitude points position of > c is different from prolate sphere, is typically in top and the bottom, 0 °≤a of longitude≤
In 45 ° and 135 °≤a≤180 °, d is obtainedab> c non-empty longitude points (Figure 15 (b)).B=φ and b=φ+180 is carried out respectively straight
Line fitting is found with the vertical range of straight line most since the actual direction of oblate spheroid has been approximate direction with least square method
The slope of small fitting a straight line:
Two fitting a straight lines are average as shown in Figure 15 (c).This direction is the longitudinal with central axis, need to be turned
90 °, and calculate △ θ:
Longitude θ=θ of central axis0+ △ θ, center axis rotation to vertical direction such as Figure 15 (d).Final tangerine orange central axis
Longitude and latitude is (94 °, 87 °), and the longitude and latitude of apple central axis is (83 °, 75 °).
4, point cloud data is handled again
(1) cloud rotation transformation is put
Cloud initial data will be put and rotate φ (central axis latitude) around Y with formula (2), rotate θ (central axis longitude) further around Z:
Point cloud when fruit center axis rotation is to vertical direction is as shown in figure 16.
(2) denoising
Using front method (1, rectangular co-ordinate switch to spherical coordinates and 2, point cloud data denoising) denoised after point cloud
Such as Figure 17.
Figure 19 is two latitudes after Figure 18 denoising through point and line chart.As can be seen from the figure a large amount of noise is eliminated
Point.
Although original point cloud data is more, not ensuring that centainly has corresponding point phase in the longitude and latitude position at setting interval
It is corresponding, there can be less amount of vacancy point, it is also seen that null point is arranged at the top of apple from Figure 19.
(3) filling of null point
Since null point is less, null point is filled according to the coordinate of non-null point within the scope of null point neighborhood longitude and latitude, using the method for average
It is filled, Size of Neighborhood takes 3 × 3 (or 5 × 5).
In formula: m is that R [a+u] [b+v] is not empty number.Figure 20 is two latitudes after Figure 19 filling null point through point
Line chart.A small amount of point is filled in the imaginary circle marked in figure.
(4) mean filter further denoises
Figure 20 through dotted line also some noises and unsmooth, finally carry out mean filter, common 3 × 3 and 5 × 5 mean values
Filter is as follows:
The center alignment of filter needs the point on the latitude coordinates (a, b) that filter, for 3 × 3 filters, the neighborhood
Latitude coordinates be (a-1, b-1), (a-1, b), (a-1, b+1), (a, b-1), (a, b+1), (a+1, b-1), (a+1, b),
The point of (a+1, b+1).
For 5 × 5 filters, which is latitude coordinates are as follows:
(a-2,b-2)、(a-2,b-1)、(a-2,b)、(a-2,b+1)、(a-2,b+2)
(a-1,b-2)、(a-1,b-1)、(a-1,b)、(a-1,b+1)、(a-1,b+2)
(a,b-2)、(a,b-1)、(a,b+1)、(a,b+2)
(a+1,b-2)、(a+1,b-1)、(a+1,b)、(a+1,b+1)、(a+1,b+2)
(a+2,b-2)、(a+2,b-1)、(a+2,b)、(a+2,b+1)、(a+2,b+2)
Figure 21 be Figure 20 after 5 × 5 mean filters through line chart.
The latitude and longitude of certain intervals (for example, 11) are taken, and are rotated by a certain angle, the grid chart of fruit such as Figure 22.
(5) fruit is rebuild
The present invention is rebuild using B-spline surface.
1) B-spline surface
B-spline surface is defined by property polyhedron, the shape approximation of the curved surface polyhedron, B-spline surface equation are as follows:
PijIt is to define polyhedral vertex, Ni,k(u) and Nj,lIt (v) is B-spline basic function.The recurrence formula of basic function defines
For (agreement 0/0=0):
Similarly
In formula, uiIt is nodal value, U=[u0,u1,…,um+k] constitute k rank B-spline knot vector, viIt is nodal value, V=
[v0,v1,…,vn+l] l rank B-spline knot vector is constituted, node is non-descending series.
B-spline surface presses the distribution situation of its knot vector interior joint, can be divided into multiple types.The present invention is using uniform
Period 2 times (k=l=3) B-spline surfaces.The curved surface is made of multiple fettucelles, has 1 rank continuous between fettucelle.It is each small
Dough sheet is controlled by the polyhedron on 9 vertex, that is, adjacent 9 mesh points in the grid body in Figure 22, formula are as follows:
In formula: Px,yFor longitude and latitude mesh point coordinate,
N0,3(t+2)=(1-t)2/2,N0,3(t+1)=- t2+t+1/2,N0,3(t)=t2/2
D is grid interval (as unit of degree)
2) BORDER PROCESSING
Borderline properties such as Figure 23 of each fettucelle of B-spline surface, it can be seen that B-spline facet sheet border and grid
There is a certain distance on boundary.Therefore, generate closed surface when, although grid body be it is closed, B-spline surface is not closed, need
Add duplicate grid.
Cavity caused by the black circular portion in the top Figure 24 is longitude grid at 180 °, but B-spline surface cannot reach 180 °.
The cavity of middle part banding pattern is caused by not repeating addition latitude grid.
The longitude points for first expanding 0 ° and 180 ° outward, enable 0 ° and 180 ° of B-spline surface longitude in:
P-1,y=2P0,y-P1,y
P180/d+1,y=2P180/d,y-P180/d-1,y
(y=0,1 ..360/d)
Then expand exhibition latitude point outward, enable B-spline surface in latitude direction closed stratum:
Px,360/d+y=Px,y(x=0,1 ... 180/d, y=1,2)
It is shone using formula (8) and plus simple optical, reconstruction figure such as Figure 25 of three fruit different angles.
It is obvious to a person skilled in the art that invention is not limited to the details of the above exemplary embodiments, Er Qie
In the case where without departing substantially from spirit or essential attributes of the invention, the present invention can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the present invention is by appended power
Benefit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent elements of the claims
Variation is included within the present invention.
Claims (7)
1. a kind of fruit point cloud method for reconstructing based on spherical coordinates, which comprises the steps of:
Step 1, based on three-dimensional fruit original point cloud data, a cloud rectangular co-ordinate is switched into spherical coordinates；
Step 2, denoising is carried out to point cloud data；
Step 3, oblate spheroid is calculated to feature when being parallel to frontal plane of projection based on oblate spheroid central axis and prolate sphere center axis rotation
The longitude and latitude of body central axis and the longitude and latitude of prolate sphere central axis；
Step 4, the longitude and latitude of longitude and latitude and prolate sphere central axis based on oblate spheroid central axis, using B-spline surface to fruit
It is rebuild.
2. a kind of fruit point cloud method for reconstructing based on spherical coordinates according to claim 1, which is characterized in that described
Step 1 includes:
In the point cloud data deposit single linked list that will acquire；
The center point coordinate for calculating point cloud recalculates a cloud coordinate deposit after point cloud center translation to rectangular co-ordinate origin
In single linked list；
A cloud rectangular co-ordinate is switched into spherical coordinates value.
3. a kind of fruit point cloud method for reconstructing based on spherical coordinates according to claim 1, which is characterized in that described
Step 2 includes:
Noise is filtered using the median filtering of sort method filter method.
4. a kind of fruit point cloud method for reconstructing based on spherical coordinates according to claim 1, which is characterized in that described
Step 3 includes:
It calculates the central point through dotted line and surrounds circle；
Calculate warp and circular difference degree c；
For a fruit, if latitude differs two of 180 ° through minimum widith w between dotted lineminCorresponding c is cmin, maximum between warp
Width wmaxCorresponding c is cmax；
Work as cmin>cmaxWhen, which is oblate spheroid, wminCorresponding latitude is the latitude φ of center axis；
Work as cmin<cmaxWhen, which is prolate sphere, wmaxCorresponding latitude is the latitude φ of center axis；
Work as wmin=wmaxWhen, central axis is in vertical direction, through dotted line height h > wminWhen, which is prolate sphere；
h<wmin
When, it is oblate spheroid；H=wminWhen be spherosome.
5. a kind of fruit point cloud method for reconstructing based on spherical coordinates according to claim 4, which is characterized in that described
Step 3 further include:
When will be through line width maximum corresponding two through dotted line when the plane is rotated to around central point through line width minimum
Longitude be determined as the initial longitude θ of prolate sphere central axis0；
By initial longitude θ0It is set as 0 °, recalculates the longitude of the longitude points of latitude b=φ, φ+180；
On the same latitude, 5 adjacent longitude points are carried out smooth；
Calculate encirclement circle of the prolate sphere through dotted line, calculate round difference degree c, calculate each non-empty longitude points and surround circle away from
From dab；
D is obtained in+180 latitude of b=φ and b=φ respectivelyabThe longitude points of > c find the level with straight line with least square method
G-bar k apart from the smallest two fitting a straight lines；
△ θ, longitude θ=θ of prolate sphere central axis are calculated based on slope k0+△θ。
6. a kind of fruit point cloud method for reconstructing based on spherical coordinates according to claim 5, which is characterized in that described
Step 3 further include:
When will be through line width minimum corresponding two through dotted line when the plane is rotated to around central point through line width maximum
Longitude be determined as the initial longitude θ of oblate spheroid central axis0；
By initial longitude θ0It is set as 0 °, recalculates the longitude of the longitude points of latitude b=φ, φ+180；
On the same latitude, 5 adjacent longitude points are carried out smooth；
Calculate encirclement circle of the oblate spheroid through dotted line, calculate round difference degree c, calculate each non-empty longitude points and surround circle away from
From dab；
Respectively in+180 latitude of b=φ and b=φ and longitude is within the scope of 0 °≤a≤45 ° and 135 °≤a≤180 °, d is obtainedab>
C non-empty longitude points are rotated by 90 ° average oblique with the smallest two fitting a straight lines of vertical range of least square method searching and straight line
Rate k；
△ θ, longitude θ=θ of oblate spheroid central axis are calculated based on slope k0+△θ。
7. a kind of fruit point cloud method for reconstructing based on spherical coordinates according to claim 1, which is characterized in that described
Step 4 includes:
Cloud rotation transformation and a denoising are carried out to original point cloud data；
Null point is filled according to the coordinate of non-null point within the scope of null point neighborhood longitude and latitude；
It is further denoised using mean filter；
Fruit reconstruction is carried out using 2 B-spline surfaces of uniform period；
The boundary of each fettucelle of B-spline surface is handled, B-spline surface is closed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811207363.7A CN109448106B (en) | 2018-10-17 | 2018-10-17 | Plant fruit point cloud reconstruction method based on spherical coordinates |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811207363.7A CN109448106B (en) | 2018-10-17 | 2018-10-17 | Plant fruit point cloud reconstruction method based on spherical coordinates |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109448106A true CN109448106A (en) | 2019-03-08 |
| CN109448106B CN109448106B (en) | 2023-03-10 |
Family
ID=65546227
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811207363.7A Active CN109448106B (en) | 2018-10-17 | 2018-10-17 | Plant fruit point cloud reconstruction method based on spherical coordinates |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109448106B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109448107A (en) * | 2018-10-17 | 2019-03-08 | 东华理工大学 | Fruit point cloud method for reconstructing based on spherical coordinates |
| CN113393514A (en) * | 2021-06-11 | 2021-09-14 | 中国科学院自动化研究所 | Three-dimensional disordered point cloud data processing method, system and equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7428476B1 (en) * | 2001-08-10 | 2008-09-23 | Yasumi Capital, Llc | System and method of simulating with respect to spheroid reference models using local surface coordinates |
| CN102915561A (en) * | 2012-09-27 | 2013-02-06 | 清华大学 | Method of three-dimensional reconstruction for pipeline structures |
| CN103077559A (en) * | 2012-12-25 | 2013-05-01 | 北京农业信息技术研究中心 | Cluster three-dimensional rebuilding method based on sequence image |
| CN103824324A (en) * | 2014-02-28 | 2014-05-28 | 北京农业信息技术研究中心 | Fruit tree crown layer leaf and fruit three-dimensional reconstruction method and system |
| CN108335354A (en) * | 2018-03-01 | 2018-07-27 | 东北林业大学 | A kind of greedy projection Triangulation Algorithm list wood method for reconstructing based on Moving Least |
| CN109448107A (en) * | 2018-10-17 | 2019-03-08 | 东华理工大学 | Fruit point cloud method for reconstructing based on spherical coordinates |
-
2018
- 2018-10-17 CN CN201811207363.7A patent/CN109448106B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7428476B1 (en) * | 2001-08-10 | 2008-09-23 | Yasumi Capital, Llc | System and method of simulating with respect to spheroid reference models using local surface coordinates |
| CN102915561A (en) * | 2012-09-27 | 2013-02-06 | 清华大学 | Method of three-dimensional reconstruction for pipeline structures |
| CN103077559A (en) * | 2012-12-25 | 2013-05-01 | 北京农业信息技术研究中心 | Cluster three-dimensional rebuilding method based on sequence image |
| CN103824324A (en) * | 2014-02-28 | 2014-05-28 | 北京农业信息技术研究中心 | Fruit tree crown layer leaf and fruit three-dimensional reconstruction method and system |
| CN108335354A (en) * | 2018-03-01 | 2018-07-27 | 东北林业大学 | A kind of greedy projection Triangulation Algorithm list wood method for reconstructing based on Moving Least |
| CN109448107A (en) * | 2018-10-17 | 2019-03-08 | 东华理工大学 | Fruit point cloud method for reconstructing based on spherical coordinates |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109448107A (en) * | 2018-10-17 | 2019-03-08 | 东华理工大学 | Fruit point cloud method for reconstructing based on spherical coordinates |
| CN109448107B (en) * | 2018-10-17 | 2023-02-24 | 东华理工大学 | Plant fruit point cloud reconstruction method based on spherical coordinates |
| CN113393514A (en) * | 2021-06-11 | 2021-09-14 | 中国科学院自动化研究所 | Three-dimensional disordered point cloud data processing method, system and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109448106B (en) | 2023-03-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Fortune | Voronoi diagrams and Delaunay triangulations | |
| Berger et al. | A survey of surface reconstruction from point clouds | |
| CN104700451B (en) | Point cloud registering method based on iterative closest point algorithm | |
| CN104778688B (en) | The method for registering and device of cloud data | |
| CN104268934B (en) | Method for reconstructing three-dimensional curve face through point cloud | |
| CN113781667A (en) | Three-dimensional structure simplified reconstruction method and device, computer equipment and storage medium | |
| CN109448106A (en) | Fruit point cloud method for reconstructing based on spherical coordinates | |
| CN109448107A (en) | Fruit point cloud method for reconstructing based on spherical coordinates | |
| CN111856747B (en) | Optimal transmission-based reflector design method | |
| CN104183021A (en) | Method for simplifying point cloud data by utilizing movable space | |
| Maurya et al. | Performance of greedy triangulation algorithm on reconstruction of coastal dune surface | |
| Tang et al. | Surface reconstruction algorithms: Review and comparison | |
| CN114155327A (en) | Intelligent multi-resolution feature optimization three-dimensional reconstruction method and system | |
| CN108665548A (en) | A kind of point cloud rapid progressive method for reconstructing assumed based on star | |
| CN105844713A (en) | Parametric surface reconstruction method based on manifold | |
| Georgiadis et al. | High quality mesh generation using cross and asterisk fields: application on coastal domains | |
| CN104346822B (en) | texture mapping method and device | |
| CN107978013B (en) | Spherical surface data organization rendering and collision detection method and system | |
| CN109979007A (en) | A kind of geometric modeling method and apparatus of building body | |
| CN114528730B (en) | Construction method of real coral sand particle discrete element model | |
| Shojaee et al. | Triangulation for surface modelling | |
| CN115984442B (en) | Method for generating surface 3D texture through feature array | |
| Phothong et al. | Generation and quality improvement of 3D models from silhouettes of 2D images | |
| Cheng | Algorithms for smooth and deformable surfaces in 3D | |
| Zhang et al. | Repair methods of complex holes in a three-dimensional model based on variation level set |
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 |