CN106056671B - A kind of contour connection of broken lines method - Google Patents
A kind of contour connection of broken lines method Download PDFInfo
- Publication number
- CN106056671B CN106056671B CN201610339589.7A CN201610339589A CN106056671B CN 106056671 B CN106056671 B CN 106056671B CN 201610339589 A CN201610339589 A CN 201610339589A CN 106056671 B CN106056671 B CN 106056671B
- Authority
- CN
- China
- Prior art keywords
- breakpoint
- pixel
- contour
- point
- initial
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T19/00—Manipulating 3D models or images for computer graphics
Landscapes
- Engineering & Computer Science (AREA)
- Computer Graphics (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of contour connection of broken lines methods.Method includes the single pixel refined image for obtaining contour broken string, using GPU algorithm identification breakpoint and connection of broken lines.Contour join algorithm of the invention had both considered the mathematical properties of contour, it is also considered that the geographical attribute of contour.It improves based on the minimum local connection method with the least commitment of direction difference of distance, during searching for match breakpoint, while the distance between breakpoint is considered and direction, provided with more robust matching rule.This method accuracy is higher, more accurately restores contour.
Description
Technical field
The invention belongs to survey and draw geography information extractive technique field, and in particular to a kind of contour connection method.
Background technique
Contour is the closed curve that the adjacent spots that elevation is equal on topographic map are linked to be, it is however generally that, contour it
Between mutually disjoint.But in the pretreatment to Pictorial Map color segmentation, during going background and color segmentation, due to depositing
The fracture that will cause contour with situations such as line feature glands is obscured in color.There are three types of concrete reasons:
First, the Pictorial Map information of scanning is imperfect, is lost part contour pixel, so that the landform before segmentation
There is fracture in figure, and then influence contour segmentation effect.This fracture has very big randomness.This causes pixel color to be distributed
Unevenly, so that certain pixel colors on contour closer to provincial characteristics color, thus by being accidentally divided into background layer,
Contour thread breakage is caused, fracture gap is possible to larger.
Second, what color segmentation mainly utilized is the colouring information of pixel, when contour is overlapped with other topographic identifications symbol
Or when intersecting, the colouring information of contour is capped, cause contour be divided to it is different divide in domain layer, such as black side
The gland of backing wire will cause the fracture in the horizontal or vertical direction of contour.
Third there is also some extra contours, i.e. auxiliary contour on topographic map is described with short dash line on the diagram.
Most current contour connection of broken lines algorithm can not solve the performance and just in the application of actual landform figure vector quantization
True rate requirement.
Sieve with it is peaceful et al. propose it is a kind of based on smallest point to the breakpoint joint method in the metallic phase image analysis of method, usual
In the case of, grain boundary of the metallographic microscope as in is approximately closed broken line, therefore, to each of image breakpoint, find out with
It is apart from nearest breakpoint and is attached thereto, and the advantages of this method is that the boundary finally obtained is unlikely to deform, but the disadvantage is that is easy
Accidentally connect breakpoint.
J Du et al. is attached the breakpoint in sparse contour region using conventional smallest point to method, other regions pass through
Select a reference curve neighbouring with breakpoint, it is desirable that matched breakpoint must satisfy and just can connect the curve is ipsilateral.Ginseng
The limitation for examining curve improves the accuracy of contour breakpoint joint well, but in the intensive region this method of contour
It is higher accidentally to connect rate, and time-consuming more.
Gul S sets a search window on the basis of range direction criterion, the direction of all breakpoints in calculation window,
And extend curve along the direction of breakpoint, using the two breakpoints as this method of matching double points if extended line intersection
The wrong even phenomenon of breakpoint can be reduced, but if two breakpoint institutes are almost parallel online possibly can not then to intersect, to cannot sentence
The breakpoint of such case is determined to although these innovatory algorithms reduce the wrong even rate of breakpoint to a certain degree, and approximation is put down
Mistake is easy between capable contour even, high thread breakage is crossed general goal and can not correctly be connected.
Li Xiao is clear et al. to combine the minimum local connection method with the least commitment of direction difference of distance, disconnected in search matching
During point, while the distance between breakpoint is considered and direction, if the distance between breakpoint is minimum and direction difference is minimum
Then connect breakpoint.This method accuracy is higher, but computation complexity is higher.
Summary of the invention
For the defects or inadequacies for existing technologies, the purpose of the present invention is to provide a kind of contour connection methods.
Contour connection of broken lines method provided by the invention includes:
(1) the single pixel refined image for obtaining contour broken string, if any pixel P in single pixel refined image0Seat
It is designated as (i, j), i takes positive integer, and j takes positive integer, for any non-P0Pixel Pr, PrCoordinate be (x, y), x takes positive integer, y
Take positive integer;
If x<I and y<J remembers PrOpposite P0Direction is direction 0;
If x=i and y<J remembers PrOpposite P0Direction is direction 1;
If x>I and y<J remembers PrOpposite P0Direction is direction 2;
If x>I and y=j remember PrOpposite P0Direction is direction 3;
If x>I and y>J remembers PrOpposite P0Direction is direction 4;
If x=i and y>J remembers PrOpposite P0Direction is direction 5;
If x<I and y>J then remembers PrOpposite P0Direction is direction 6;
If x<I and y=j, then remember PrOpposite P0Direction is direction 7;
Wherein, direction 0 and direction 4 mutual opposite direction each other, direction 2 and direction 6 mutual opposite direction each other, side
To 0 and the mutual parabola direction each other of direction 2, direction 2 and direction 4 mutual parabola direction each other, direction 4 and direction 6
Mutual opposite direction each other, direction 6 and direction 0 mutual opposite direction each other;
(2) breakpoint is identified using GPU algorithm
Using CUDA threading model by the pixel one-to-one mapping on multiple GPU threads and single pixel refined image, often
A GPU thread executes following steps:
(2.1) initial breakpoint is detected
For the pixel of current thread mapping, if the pixel meets formula (1) condition, which is initial disconnected
Point:
V0For the pixel value of the corresponding pixel of current thread, V1、V2、V3、V4、V5、V6、V7And V8Respectively pixel V0
Eight field pixels pixel value;
(2.2) breakpoint match
(a) the initial breakpoint P to be detected in step (2.1)1Centered on, search meets formula in window size W area
(1) breakpoint searched is added in Candidate Set S the breakpoint of condition shown in, and the window size W meets:10 > W > 5, and
W takes positive integer;
(b) the part breakpoint in point set S is deleted, obtains point set S', the direction of each breakpoint in point set S' and initial breakpoint
P1Direction each other opposite direction or with initial breakpoint P1Direction parabola direction each other;
The direction definition of the breakpoint is:For the break-point in Candidate Set S, returned along contour where the break-point
Multiple pixels trace back to a certain backtracking pixel, relative to the backtracking pixel, which is located at direction 0, direction 1, side
To 2, direction 3, direction 4, direction 5, direction 6 or direction 7, which is this relative to the backtracking pixel direction
The direction of meaning breakpoint;
(c) delete point set S' in initial breakpoint P1Euclidean distance DsMore than distance threshold δ1Breakpoint, wherein 10 >
δ1> 8 deletes the breakpoint P for meeting formula (3) in point set S'a, Pa∈ S' obtains point set S ";
max{|θ(Pa,La)-θ(Pa,P1)|,|θ(P1,L1)-θ(Pa,P1) | < δ2 (3)
In formula (3):
LaFor breakpoint PaPlace contour,
L1For initial breakpoint P1Place contour,
θ(Pa,P1) it is breakpoint PaWith initial breakpoint P1Tilt angle of the straight line relative to abscissa where two o'clock,
θ(P1,L1) it is contour L1Relative to the inclination angle of abscissa,m
=2,3 ..., k1, initial breakpoint P1Along contour L1M-1 pixel is recalled to pixel Pm, 10 > k1> 5, k1It takes just whole
Number;
θ(Pa,La) it is contour LaRelative to the inclination angle of abscissa,n
=2,3 ..., k2, breakpoint PaAlong contour LaN-1 pixel is recalled to pixel Pn, 10 > k2> 5, k2For positive integer;
(d) each Candidate point and initial breakpoint P in point set S " are sought1Metric function, wherein minimum metric functional value
Corresponding Candidate point is initial breakpoint P1Optimal selection point;Wherein:Breakpoint P1,PbMetric function be defined as:
In formula (4):θ(P1,L1) it is contour L1Relative to the tilt angle of abscissa, θ (Pb,Lb) it is contour LbRelatively
In the tilt angle of abscissa, θ (P1,Pb) it is breakpoint PbWith initial breakpoint P1Inclination of the straight line relative to abscissa where two o'clock
Angle, φ1+φ2=1, φ1And φ2It is positive real number, Ds(P1,Pb) it is breakpoint P1And PbBetween Euclidean distance;
(3) connection of broken lines
After the completion of the execution of all threads, connect initial breakpoint and its optimal selection point one by one, such as a certain initial breakpoint with
There are crosspoints with existing contour for line between optimal selection point, then delete the optimal selection point and initial breakpoint.
Further, for the pixel of current thread mapping in step (2.1), if the pixel meets formula (2) condition,
Then the pixel is crosspoint:
V0For the pixel value of the corresponding pixel of current thread, V1、V2、V3、V4、V5、V6、V7And V8Respectively pixel V0
Eight field pixels pixel value.
Further, the direction definition of step (2.2) described breakpoint is:It is any along this for the break-point in Candidate Set S
Contour where breakpoint carries out at least ten pixels of backtracking to a certain backtracking pixel, relative to the backtracking pixel, this
Meaning breakpoint is located at direction 0, direction 1, direction 2, direction 3, direction 4, direction 5, direction 6 or direction 7, and the break-point is relative to this
Recall the direction that pixel direction is the break-point.
Compared with prior art, the present invention has the following advantages that:
First, contour join algorithm proposed by the present invention had both considered the mathematical properties of contour, it is also considered that contour
The geographical attribute of line.It improves based on the minimum local connection method with the least commitment of direction difference of distance, it is disconnected in search matching
During point, while the distance between breakpoint is considered and direction, provided with more robust matching rule.This method accuracy
It is higher, contour is restored more accurately.
Second, the present invention is when carrying out the recovery of contour thread breakage, and using GPU video card, the mode of integrating parallel programming will break
Line join algorithm optimization, so that algorithm improves the real-time of algorithm while keeping high-accuracy.Compared to primal algorithm,
Algorithm accelerates 1,000 times or more, meets the needs of Pictorial Map vector quantization application.
Third, the present invention be applied to color segmentation after brown version contour fault rupture reconstruction processing, can effectively remove by
The high thread breakage such as color segmentation bring and square rid gland lead to contour deficient phenomena, and can restore landform out well
The contour information lost in figure.
Detailed description of the invention
Fig. 1 is eight directions of opposite central pixel point;
Fig. 2 is the eight neighborhood pixel of any pixel;
Fig. 3 is the simulation experiment result.
Specific embodiment
Contour is the closed curve that the adjacent spots that elevation is equal on topographic map are linked to be, it is however generally that, contour it
Between mutually disjoint.But in the pretreatment to Pictorial Map color segmentation, during going background and color segmentation, due to depositing
The fracture that will cause contour with situations such as line feature glands is obscured in color.There are three types of concrete reasons:First, the colour of scanning
Information of the ground map is imperfect, is lost part contour pixel, so that the topographic map before segmentation has fracture, and then influences contour
Line segmentation effect.This fracture has very big randomness.This causes pixel color to be unevenly distributed, so that certain on contour
A little pixel colors closer to provincial characteristics color, to cause contour thread breakage, faulted joint by being accidentally divided into background layer
Gap is possible to larger.Second, what color segmentation mainly utilized is the colouring information of pixel, when contour and other topographic identifications accord with
When being overlapped or intersecting, the colouring information of contour is capped, cause contour be divided to it is different divide in domain layer, it is such as black
The gland of color square rid will cause the fracture in the horizontal or vertical direction of contour.Third, there is also some extra contours on topographic map,
That is auxiliary contour is described with short dash line on the diagram.
To going the Pictorial Map after background to carry out color segmentation, the contour hierarchical diagram of brown version is obtained.In contour
Before fracture connection, it is also necessary to carry out pre-processing to brown version contour map picture, mainly include filtering and noise reduction, binaryzation, refinement
With go right angle etc. operate, finally obtain the contour refined image of the wide fracture of single pixel, method of the invention is to realize
The connection of high line broken string, is further explained explanation to technical solution of the present invention with reference to the accompanying drawing.
Step 1, the single pixel refined image of contour broken string is obtained
Referring to Fig.1, if arbitrary initial pixel P in single pixel refined image0Coordinate be (i, j), i takes positive integer, and j takes
Positive integer, for any non-P0Pixel Pr, PrCoordinate be (x, y), x takes positive integer, and y takes positive integer;
If x<I and y<J remembers PrOpposite P0Direction is direction 0;
If x=i and y<J remembers PrOpposite P0Direction is direction 1;
If x>I and y<J remembers PrOpposite P0Direction is direction 2;
If x>I and y=j remember PrOpposite P0Direction is direction 3;
If x>I and y>J remembers PrOpposite P0Direction is direction 4;
If x=i and y>J remembers PrOpposite P0Direction is direction 5;
If x<I and y>J remembers PrOpposite P0Direction is direction 6;
If x<I and y=j remember PrOpposite P0Direction is direction 7;
Wherein, direction 0 and direction 4 mutual opposite direction each other, direction 2 and direction 6 mutual opposite direction each other, side
To 0 and the mutual parabola direction each other of direction 2, direction 2 and direction 4 mutual parabola direction each other, direction 4 and direction 6
Mutual opposite direction each other, direction 6 and direction 0 mutual opposite direction each other;
Step 2, using the detection of GPU algorithm and match breakpoint
Using CUDA threading model by the pixel one-to-one mapping on multiple GPU threads and single pixel refined image, often
A GPU thread executes following steps:
2a) detect initial breakpoint
Referring to Fig. 2, pixel corresponding for current thread, if the pixel meets formula (1) condition, the pixel
For initial breakpoint, if the pixel meets formula (2) condition, which is crosspoint:
V0For the pixel value of the corresponding pixel of current thread, V1、V2、V3、V4、V5、V6、V7And V8Respectively pixel V0
8 neighborhood territory pixel points pixel value;
2b) breakpoint match
(1) the initial breakpoint P to be detected in step 2a)1Centered on, other breakpoints are searched in window size W area,
Each breakpoint is all satisfied condition shown in formula (1), and the breakpoint searched is added in Candidate Set S, and the window size W meets:
10 > W > 5, and W takes positive integer, W takes 8 in the embodiment;
(2) the part breakpoint in point set S is deleted, obtains point set S', the direction of each breakpoint in point set S' and initial breakpoint
P1Direction each other opposite direction or with initial breakpoint P1Direction parabola direction each other;
For the break-point in Candidate Set S, at least ten pixels of backtracking are carried out along contour where the break-point
To a certain backtracking pixel, relative to the backtracking pixel, break-point is located at direction 0, direction 1, direction 2, direction 3, direction
4, direction 5, direction 6 or direction 7, the break-point direction are the direction of the break-point;
(3) delete point set S' in initial breakpoint P1Euclidean distance DsMore than distance threshold δ1Breakpoint, wherein 10 >
δ1> 8, Pa∈ S', 20 ° of > δ20 ° of >;Delete the breakpoint P for meeting formula (3) in point set S'a, obtain point set S ";δ in the embodiment1
It is 8.5, δ2It is 18 °.
max{|θ(Pa,La)-θ(Pa,P1)|,|θ(P1,L1)-θ(Pa,P1) | < δ2 (3)
In formula (3), LaFor breakpoint PaPlace contour, L1For initial breakpoint P1Place contour, θ (Pa,P1) it is breakpoint Pa
With initial breakpoint P1Tilt angle of the straight line relative to abscissa where two o'clock,
θ(P1,L1) it is contour L1Relative to the inclination angle of abscissa,m
=2,3 ..., k1, initial breakpoint P1Along contour L1M-1 pixel is recalled to pixel Pm, 10 > k1> 5, k1It takes just whole
Number;
θ(Pa,La) it is contour LaRelative to the inclination angle of abscissa,n
=2,3 ..., k2, breakpoint PaAlong contour LaN-1 pixel is recalled to pixel Pn, 10 > k2> 5, k2For positive integer;
(4) each Candidate point and initial breakpoint P in point set S " are sought1Metric function, wherein minimum metric functional value
Corresponding Candidate point is breakpoint P1Optimal selection point;Any two breakpoint P1,PbMetric function be defined as:
φ(P1,Pb)=max | θ (P1,L1)-θ(P1,Pb) |, | θ (Pb,Lb)-θ(P1,Pb)}×φ1+Ds(P1,Pb)×φ2
Wherein, θ (P1,L1) it is contour L1Relative to the tilt angle of abscissa, θ (Pb,Lb) it is contour LbRelative to
The tilt angle of abscissa, acquiring method refer to θ (P1,L1), θ (P1,Pb) it is breakpoint PbWith initial breakpoint P1Straight line where two o'clock
Tilt angle relative to abscissa;φ1And φ2Respectively angle weight and apart from weight, and meet:φ1+φ2=1, φ1With
φ2It is positive real number, Ds(P1,Pb) it is breakpoint P1And PbBetween Euclidean distance;φ in the embodiment1It is 0.45, φ2It is 0.55.
Step 3, connection of broken lines
After the completion of the execution of all threads, connect initial breakpoint and its optimal selection point one by one, such as a certain initial breakpoint with
There are crosspoints with existing contour for line between optimal selection point, then delete the optimal selection point
With initial breakpoint, the line finally retained is contour.
Emulation connection is carried out to sample according to above-described embodiment, the topographic contour that simulation sample is 6600x8000 refines
The connection effect of image, the video card to different model realized under OpenCV 2.4.9 and CUDA C programmed environment carries out
Comparison.Fig. 3 is contour connection of broken lines label figure, and the join domain of contour is marked in thick lines part in figure.It is tied from emulation
For fruit it is observed that most of contour thread breakage can restore, the case where incorrect link, is less.Simultaneously as GPU video card
Accelerate, the real-time of algorithm is fully reflected.Different video card contour connection effect comparisons are as shown in table 1 below.
The different video card contour connection effect comparisons of table 1
It can see from table, using common 760 video card of NVIDIA GEFORCE GTX to 6600x8000 topographic map
Contour connection of broken lines is carried out, the used time about 5 seconds, accelerates about 816 times, using the calculation type video card NVIDIA Tesla of mainstream
K40,1360 times of the contour connection of broken lines algorithm acceleration based on GPU acceleration, while algorithm maintain consistent and stable with CPU
Bonding ratio, bonding ratio be have connected breakpoint number with connect before the ratio between breakpoint number.
Claims (3)
1. a kind of contour connection of broken lines method, which is characterized in that method includes:
(1) the single pixel refined image for obtaining contour broken string, if any pixel P in single pixel refined image0Coordinate be
(i, j), i take positive integer, and j takes positive integer, for any non-P0Pixel Pr, PrCoordinate be (x, y), x takes positive integer, and y takes just
Integer;
If x<I and y<J remembers PrOpposite P0Direction is direction 0;
If x=i and y<J remembers PrOpposite P0Direction is direction 1;
If x>I and y<J remembers PrOpposite P0Direction is direction 2;
If x>I and y=j remember PrOpposite P0Direction is direction 3;
If x>I and y>J remembers PrOpposite P0Direction is direction 4;
If x=i and y>J remembers PrOpposite P0Direction is direction 5;
If x<I and y>J then remembers PrOpposite P0Direction is direction 6;
If x<I and y=j, then remember PrOpposite P0Direction is direction 7;
Wherein, direction 0 and direction 4 mutual opposite direction each other, direction 2 and direction 6 mutual opposite direction each other, direction 0
With the mutual parabola direction each other of direction 2, direction 2 and direction 4 mutual parabola direction each other, direction 4 and direction 6 are each other
Mutual opposite direction, direction 6 and direction 0 mutual opposite direction each other;
(2) breakpoint is identified using GPU algorithm
Using CUDA threading model by the pixel one-to-one mapping in multiple GPU threads and single pixel refined image, each GPU
Thread executes following steps:
(2.1) initial breakpoint is detected
For the pixel of current thread mapping, if the pixel meets formula (1) condition, which is initial breakpoint:
V0For the pixel value of the corresponding pixel of current thread, V1、V2、V3、V4、V5、V6、V7And V8Respectively pixel V0It is eight adjacent
The pixel value of domain pixel;
(2.2) breakpoint match
(a) the initial breakpoint P to be detected in step (2.1)1Centered on, search meets formula (1) institute in window size W area
The breakpoint searched is added in Candidate Set S the breakpoint for showing condition, and the window size W meets:10>W>5, and W take it is just whole
Number;
(b) the part breakpoint in point set S is deleted, point set S' is obtained, the direction of each breakpoint in point set S' and initial breakpoint P1Side
To opposite direction each other or with initial breakpoint P1Direction parabola direction each other;
The direction definition of the breakpoint is:It is more along contour backtracking where the break-point for the break-point in Candidate Set S
A pixel is to a certain backtracking pixel, relative to the backtracking pixel, the break-point be located at direction 0, direction 1, direction 2,
Direction 3, direction 4, direction 5, direction 6 or direction 7, the break-point are that this is any disconnected relative to the backtracking pixel direction
The direction of point;
(c) delete point set S' in initial breakpoint P1Euclidean distance DsMore than distance threshold δ1Breakpoint, wherein 10>δ1>8, it deletes
Except the breakpoint P for meeting formula (3) in point set S'a, Pa∈ S' obtains point set S ";
max{θ(Pa,La)-θ(Pa,P1)|,|θ(P1,L1)-θ(Pa,P1)|}<δ2 (3)
In formula (3):
LaFor breakpoint PaPlace contour,
L1For initial breakpoint P1Place contour,
θ(Pa,P1) it is breakpoint PaWith initial breakpoint P1Tilt angle of the straight line relative to abscissa where two o'clock,
20°>δ2>0 °,
θ(P1,L1) it is contour L1Relative to the inclination angle of abscissa,M=2,
3,...,k1, initial breakpoint P1Along contour L1M-1 pixel is recalled to pixel Pm, 10>k1>5, k1Take positive integer;
θ(Pa,La) it is contour LaRelative to the inclination angle of abscissa,N=2,
3,...,k2, breakpoint PaAlong contour LaN-1 pixel is recalled to pixel Pn, 10>k2>5, k2For positive integer;
(d) each Candidate point and initial breakpoint P in point set S " are sought1Metric function, wherein minimum metric functional value is corresponding
Candidate point is initial breakpoint P1Optimal selection point;Wherein:Breakpoint P1,PbMetric function be defined as:
In formula (4):θ(P1,L1) it is contour L1Relative to the tilt angle of abscissa, θ (Pb,Lb) it is contour LbRelative to cross
The tilt angle of coordinate, θ (P1,Pb) it is breakpoint PbWith initial breakpoint P1Tilt angle of the straight line relative to abscissa where two o'clock,
φ1And φ2Respectively angle weight and apart from weight, φ1+φ2=1, φ1And φ2It is positive real number, Ds(P1,Pb) it is breakpoint P1
And PbBetween Euclidean distance;
(3) connection of broken lines
After the completion of the execution of all threads, connect initial breakpoint and its optimal selection point one by one, such as a certain initial breakpoint with it is optimal
There are crosspoints with existing contour for line between selected element, then delete the optimal selection point and initial breakpoint.
2. contour connection of broken lines method as described in claim 1, which is characterized in that for current thread in step (2.1)
The pixel of mapping, if the pixel meets formula (2) condition, which is crosspoint:
V0For the pixel value of the corresponding pixel of current thread, V1、V2、V3、V4、V5、V6、V7And V8Respectively pixel V0It is eight adjacent
The pixel value of domain pixel.
3. contour connection of broken lines method as described in claim 1, which is characterized in that the direction of step (2.2) described breakpoint
It is defined as:For the break-point in Candidate Set S, at least ten pixels of backtracking are carried out extremely along contour where the break-point
A certain backtracking pixel, relative to the backtracking pixel, which is located at direction 0, direction 1, direction 2, direction 3, direction
4, direction 5, direction 6 or direction 7, the break-point are the direction of the break-point relative to the backtracking pixel direction.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610339589.7A CN106056671B (en) | 2016-05-20 | 2016-05-20 | A kind of contour connection of broken lines method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610339589.7A CN106056671B (en) | 2016-05-20 | 2016-05-20 | A kind of contour connection of broken lines method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106056671A CN106056671A (en) | 2016-10-26 |
CN106056671B true CN106056671B (en) | 2018-11-16 |
Family
ID=57176559
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610339589.7A Active CN106056671B (en) | 2016-05-20 | 2016-05-20 | A kind of contour connection of broken lines method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106056671B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114037682A (en) * | 2021-11-08 | 2022-02-11 | 中国科学院光电技术研究所 | Two-dimensional automatic detection method for optical element surface defects |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0422245A1 (en) * | 1989-04-04 | 1991-04-17 | Ntt Data Communications Systems Corporation | Recognizing apparatus |
CN1647138A (en) * | 2002-04-22 | 2005-07-27 | Dgs计算机株式会社 | Digital altimetric map drawing method and device |
CN101770570A (en) * | 2009-01-04 | 2010-07-07 | 茂晖科技股份有限公司 | High-efficiency fingerprint image processing method |
CN104809692A (en) * | 2015-05-12 | 2015-07-29 | 山东临沂烟草有限公司 | Interpolation method for DEM (Digital Elevation Model) generated by contour lines |
-
2016
- 2016-05-20 CN CN201610339589.7A patent/CN106056671B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0422245A1 (en) * | 1989-04-04 | 1991-04-17 | Ntt Data Communications Systems Corporation | Recognizing apparatus |
CN1647138A (en) * | 2002-04-22 | 2005-07-27 | Dgs计算机株式会社 | Digital altimetric map drawing method and device |
CN101770570A (en) * | 2009-01-04 | 2010-07-07 | 茂晖科技股份有限公司 | High-efficiency fingerprint image processing method |
CN104809692A (en) * | 2015-05-12 | 2015-07-29 | 山东临沂烟草有限公司 | Interpolation method for DEM (Digital Elevation Model) generated by contour lines |
Non-Patent Citations (6)
Title |
---|
"Advanced contour reconnection in scanned topographic maps";Emrah Hancer等;《2011 5th International Conference on Application of Information and Communication Technologies (AICT)》;20111014;1-5 * |
"一种改进的断裂等高线重建方法研究";高帅等;《地理信息空间》;20150619;第13卷(第3期);52-55 * |
"几种CUDA加速高斯滤波算法的比较";刘进锋;《计算机工程与应用》;20130729;第49卷(第23期);14-18 * |
"基于断点分类和方向差异的等高线分级补断方法";李晓宁等;《四川大学学报 自然科学版》;20130728;第50卷(第4期);737-742 * |
"彩色地形图中等高线断裂重建算法的研究";李晓清;《中国优秀硕士学位论文全文数据库 信息科技辑》;20160315(第03期);I138-6693 * |
"栅格地形图断裂等高线的自动重建";王峰等;《计算机辅助设计与图形学学报》;20150815;第27卷(第8期);1434-1441 * |
Also Published As
Publication number | Publication date |
---|---|
CN106056671A (en) | 2016-10-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Hu et al. | Revisiting single image depth estimation: Toward higher resolution maps with accurate object boundaries | |
Ji et al. | Vehicle detection in remote sensing images leveraging on simultaneous super-resolution | |
WO2021249233A1 (en) | Image processing method, target recognition model training method, and target recognition method | |
JP4928310B2 (en) | License plate recognition device, control method thereof, computer program | |
Wei et al. | A coarse-to-fine multi-stream hybrid deraining network for single image deraining | |
CN109784386A (en) | A method of it is detected with semantic segmentation helpers | |
CN102750701B (en) | Method for detecting spissatus and spissatus shadow based on Landsat thematic mapper (TM) images and Landsat enhanced thematic mapper (ETM) images | |
CN108492248A (en) | Depth map super-resolution method based on deep learning | |
CN109858325B (en) | Table detection method and device | |
US8687895B2 (en) | Image processing apparatus, image processing method, and computer-readable medium | |
Roy et al. | Wavelet-gradient-fusion for video text binarization | |
CN109712071A (en) | Unmanned plane image mosaic and localization method based on track constraint | |
CN110399873A (en) | ID Card Image acquisition methods, device, electronic equipment and storage medium | |
CN109741325A (en) | A kind of wiring verticality intelligent detecting method | |
CN106056671B (en) | A kind of contour connection of broken lines method | |
Weng et al. | Rotational rectification network: enabling pedestrian detection for mobile vision | |
CN103871089B (en) | Image superpixel meshing method based on fusion | |
Li et al. | Multi-discriminator generative adversarial network for high resolution gray-scale satellite image colorization | |
Fan et al. | A run-length-coding-based approach to stroke extraction of Chinese characters | |
Chen et al. | Coupled global–local object detection for large vhr aerial images | |
Grm et al. | Face hallucination revisited: An exploratory study on dataset bias | |
CN105740818A (en) | Artificial mark detection method applied to augmented reality | |
Cordella et al. | Writing order recovery from off-line handwriting by graph traversal | |
Henderson et al. | Raster map image analysis | |
Han et al. | Segmentation Is Not the End of Road Extraction: An All-Visible Denoising Autoencoder for Connected and Smooth Road Reconstruction |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20210121 Address after: Building 28 and 29, Tian'an Digital City, No.88 Chunyang Road, Chengyang District, Qingdao City, Shandong Province Patentee after: Qingdao Institute of computing technology Xi'an University of Electronic Science and technology Address before: 710071 No. 2 Taibai South Road, Shaanxi, Xi'an Patentee before: XIDIAN University |