CN109785371A - A kind of sun image method for registering based on normalized crosscorrelation and SIFT - Google Patents
A kind of sun image method for registering based on normalized crosscorrelation and SIFT Download PDFInfo
- Publication number
- CN109785371A CN109785371A CN201811558754.3A CN201811558754A CN109785371A CN 109785371 A CN109785371 A CN 109785371A CN 201811558754 A CN201811558754 A CN 201811558754A CN 109785371 A CN109785371 A CN 109785371A
- Authority
- CN
- China
- Prior art keywords
- image
- point
- key point
- sun
- scale
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 230000009466 transformation Effects 0.000 claims abstract description 5
- 230000000007 visual effect Effects 0.000 claims abstract description 5
- 238000001514 detection method Methods 0.000 claims abstract description 3
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims description 18
- 238000005070 sampling Methods 0.000 claims description 7
- 239000011159 matrix material Substances 0.000 claims description 6
- BBEAQIROQSPTKN-UHFFFAOYSA-N pyrene Chemical compound C1=CC=C2C=CC3=CC=CC4=CC=C1C2=C43 BBEAQIROQSPTKN-UHFFFAOYSA-N 0.000 claims description 6
- 230000002708 enhancing effect Effects 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000000694 effects Effects 0.000 claims description 3
- GVEPBJHOBDJJJI-UHFFFAOYSA-N fluoranthrene Natural products C1=CC(C2=CC=CC=C22)=C3C2=CC=CC3=C1 GVEPBJHOBDJJJI-UHFFFAOYSA-N 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 3
- 239000000203 mixture Substances 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000013519 translation Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 abstract description 2
- 238000012545 processing Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 238000005314 correlation function Methods 0.000 description 2
- 206010064127 Solar lentigo Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of sun image method for registering based on normalized crosscorrelation and SIFT, belongs to astronomical technology and field of image processing.The present invention carries out down-sampled pretreatment respectively from the full-time face observed image of space base (SDO) and the sun local high resolution observed image of ground-based telescope (NVST) to one group first, then normalized crosscorrelation matching algorithm is used, solve visual field inconsistence problems in region subject to registration, the best match position for calculating full-time face image and Regional High Resolution image intercepts subgraph as reference picture to be registered using the position；Feature detection is carried out to image subject to registration using the feature detective operators (SIFT) based on Scale invariant, obtains feature point set；Error hiding characteristic point pair is eliminated using MLESAC algorithm；The transformation parameter between image finally is solved using least square method, obtains registration result.The present invention solve different observation sources sun image fast and automatically, high registration accuracy.
Description
Technical field
The present invention relates to a kind of sun image method for registering based on normalized crosscorrelation and SIFT, belong to astronomical technology and
Field of image processing.
Background technique
Sun observation is the important research content of Solar Physics research, main by developing various horizontal solar telescopes
The various physical phenomenons and process that occur on the sun are observed, and analysis is carried out to observation data and extracts excavation to obtain
Research achievement.Ground horizontal solar telescope and space base horizontal solar telescope two major classes are divided into helioscope.China is completed
And what is come into operation has Beijing Huairou ground telescope, 1 meter of Yunnan Observatory Fuxian Lake new vacuum infrared horizontal solar telescope
(NVST).The above are ground-based optical telescopes, can only receive the solar radiation greater than 290nm, and other wave band observation data are big
Gas-bearing formation absorbs.Meanwhile China is also completed radio telescope and is used to receive the observation letter that atmosphere absorbs its weaker all band
Number.Preferably to carry out multiwave sight to the sun, several space base observatories are had been established in the U.S. and European Union, wherein with SDO
(Solar Dynamics Observatory, SDO), Stereo and SOHO are most representative, realize to sun multiband, full-time
Domain observation.SDO carries 3 sun observation instrument, respectively obtains ultraviolet observed image on each wave band such as extreme ultraviolet.Space borne detection
It is limited to the limitation of aircraft throw-weight, full-time face observed image is difficult to realize the knot that the high-resolution as obtained ground is observed
Fruit, such as the resolution sun sunspot high-resolution observed image of NVST and NST.New observation method and means are explored, is disclosed too
Positive internal nature rule, is always the crucial research contents in sun observation.
Summary of the invention
Match the technical problem to be solved by the present invention is providing a kind of sun image based on normalized crosscorrelation and SIFT
Quasi- method, to solve the full-time face sun observation image of space base telescope shooting and the sun part height of ground telescope shooting
The problem of resolution ratio observed image is mutually registrated.
The technical solution adopted by the present invention is that: a kind of sun image method for registering based on normalized crosscorrelation and SIFT,
Include the following steps:
Step 1: two width sun observation images of input, wherein a width is the full-time face observed image from SDO, as reference
Image I1, another width is the sun local high resolution observed image from NVST, as floating image I2, to two images point
It is not pre-processed, down-sampled to two images progress first, the full-time figure of SDO reduces 0.5 times, NVST local high resolution image
Reduce 0.1 times.Using rayleigh distributed, enhancing coefficient is 0.2 couple of two images I1、I2Limitation contrast histogram equalization is carried out to increase
By force.
Step 2: using normalized crosscorrelation matching algorithm, solve visual field inconsistence problems in region subject to registration, calculate full-time
The best match position of face image and Regional High Resolution image intercepts subgraph as reference picture to be registered using the position.
The principle of normalized crosscorrelation matching algorithm is: the size of image to be detected S is M1×M2, the size of template image T
For N1×N2, general M > > N, i.e. M1Greater than N1、M2Greater than N2, template image T on image to be detected S from left to right, on to
Under search for pixel-by-pixel, the subgraph that search window is covered is denoted as SI, j, wherein (i, j) is the top left corner apex of subgraph to be detected
Scheme the coordinate in S.The gray scale related coefficient between each subgraph and Prototype drawing is calculated by correlation function R (i, j), coefficient is maximum
Subgraph coordinate be matching position.
Step 3: being detected and matched with SIFT feature.
Step3.1: establishing scale space, detects key point:
The scale space function L (x, y, σ) of one sub-picture I (x, y) can be defined as making image the Gauss of variable dimension
Convolution:
L (x, y, σ)=G (x, y, σ) * I (x, y)
Wherein G (x, γ, σ) is dimensional Gaussian kernel function, is indicated are as follows:
In order to effectively detect stable key point in scale space, need by difference of Gaussian DoG
(Difference of Gaussians) constructs difference of Gaussian pyramid, that is, utilizes the Gaussian difference pyrene and image of different scale
Carry out convolution:
D (x, y, σ)=(G (x, y, k σ)-G (x, y, σ)) * (x, y)=L (x, y, k σ)-L (x, y, σ)
Key point is made of the Local Extremum in the space DoG, and tentatively detecting for key point is by each in same group
Compare completion between DoG adjacent two layers image.In order to find DoG Function Extreme Value point, each pixel is all with it
Consecutive points compare, see whether it bigger than the consecutive points of its image area and scale domain or small.Intermediate test point is same with it
Totally 26 points compare for 8 consecutive points of scale and corresponding 9 × 2 points of neighbouring scale, to ensure in scale space and two
Dimension image space all detects extreme point.
There is stronger skirt response since DoG value is more sensitive to noise and edge, and in image border.It is detected to above-mentioned
Extreme point eliminate edge effect, and remove low-contrast point, so that it may obtain more accurate extreme point namely key point.
Step3.2: the distribution of key point direction generates key point description:
The gradient for seeking each extreme point assigns direction according to the peak value of histogram of gradients for extreme point.
The gradient of pixel indicates:
Gradient magnitude:
Gradient direction:
To image-region piecemeal around key point, calculation block inside gradient histogram generates unique vector, this
Vector is that one kind of the regional image information is abstract, has uniqueness.
Step3.3: key point matching:
Subclass is described to key point is established with reference to figure and floating figure, calculates similarity next of description of two set
With key point.In the case where characteristic point is few, the matching of key point directlys adopt the method for exhaustion to complete.
With reference to key point description in figure:
Ri=(ri1,ri2,L,ri128,)
Key point description in floating figure:
Si=(si1,si2,L,si128,)
Any two describe sub- similarity measurement:
Key point description matched, d (Ri,Si) need to meet:
Critical value Threshold is 0.6-0.75；
Step 4:MLESAC algorithm rejects Mismatching point.
MLESAC algorithm is to introduce Mind on statistics on the basis of random sampling consistency (RANSAC) algorithm, pass through
The innovatory algorithm that likelihood score is assessed the model that generates to random sampling.
MLESAC algorithm indicates that error is distributed using mixture probabilistic model:
Wherein e is evaluated error, and σ is Gaussian Profile variance, and γ is probabilistic model weighted value, and v is equally distributed search
Window constant.
It can be seen that interior point distribution using Gaussian Profile from the probabilistic model of above formula, and exterior point (outliers) is adopted
Then it is uniformly distributed.And the method that MLESAC algorithm still uses random sampling, such a solution maximal possibility estimation
Problem, which can be converted to, solves cost function minimum problems.Cost function is as shown in formula.
Step 5: transformation parameter estimation.
If some coordinates are (x0,y0), then coordinate (the x after affine transformation1,y1) can be expressed in matrix as:
[x1 y11]=[x0 y0 1]×T
Wherein T is affine matrix:
Wherein t11,t12,t21,t22It is rotation and zoom scale parameter, and tx,tyIt is translation parameters.Simply by the presence of 3 couples of spies
Levy point, so that it may which simultaneous equations solve affine parameter；It is asked when there are 4 pairs or more numbers to characteristic point using least square method
Solve more accurate affine parameter.
Further, if the affine parameter that step 5 solves after executing once does not reach requirement accurately, by step 5
Execute it is primary after obtained registration result as floating figure repeat the above steps 3,4,5, after the wheel of iteration 3 to 5, obtain high-precision
Affine parameter.
The beneficial effects of the present invention are: for the sun local high resolution observed image and space base of the shooting of ground telescope
The registration problems of the full-time face observed image of telescope shooting, propose a kind of fast automatic method for registering.It is based on compared to other
The method for registering of cross-correlation method does not need to provide precompensation parameter in advance.Arbitrarily two images subject to registration of input, it is mutual using normalization
Related algorithm can realize Automatic-searching region to be matched, as with reference to figure, will be greatly reduced in this way under the interception of the region
The calculation amount of registration, reduces the time overhead of registration.In addition it on the basis of traditional SIFT algorithm characteristics point is matched, uses
MLESAC has screened out Mismatching point, and registration accuracy improves.Algorithm does not need to iterate during realizing, disposable to be registrated
As a result sub-pixed mapping rank is had reached, is roughly equal to 0.18 rad, precision alreadys exceed the methods of mutual information and cross-correlation and iterates
Out as a result, efficiency of algorithm greatly improves.On the other hand, in the case where required precision is more harsh, this method can also be with
Continue number of iterations wheel, experimental result is shown, when iteration 3 to 5 is taken turns, precision reaches highest.In contrast, the registration efficiency of this method
It is very high.
Detailed description of the invention
Fig. 1 is flow chart of the method for the present invention；
Fig. 2 is the full-time face observed image of SDO；
Fig. 3 is NVST sun local high resolution observed image；
Fig. 4 is the region subject to registration that normalized crosscorrelation is matched to；
Fig. 5 is the matching double points that SIFT algorithm detects；
Fig. 6 is that MLESAC removes the matching double points after error hiding；
Fig. 7 is registration result figure；
Fig. 8 be iteration repeatedly after match point logarithm variation diagram；
Fig. 9 be iteration repeatedly after root-mean-square error variation diagram；
Figure 10 is the registration result precision figure of any multi-group data.
Specific embodiment
In the following with reference to the drawings and specific embodiments, the invention will be further described.
Embodiment 1: as Figure 1-10 shows, in order to prove the generality of this method, four width difference is chosen in the present embodiment
Image compare explanation.
A kind of sun image method for registering based on normalized crosscorrelation and SIFT, includes the following steps:
Step 1: two width sun observation images of input, wherein a width is the full-time face observed image from SDO (such as Fig. 2 institute
Show) as reference picture I1, another width is the sun local high resolution observed image (as shown in Figure 3) from NVST, as
Floating image I2, two images are pre-processed respectively, first two images are carried out with down-sampled, the full-time figure diminution 0.5 of SDO
Times, 0.1 times of NVST local high resolution image down.Using rayleigh distributed, enhancing coefficient is 0.2 couple of two images I1、I2It carries out
Limit the enhancing of contrast histogram equalization.
Step 2: using normalized crosscorrelation matching algorithm, solve visual field inconsistence problems in region subject to registration, calculate full-time
The best match position (as shown in Figure 4) of face image and Regional High Resolution image intercepts subgraph as to be registered using the position
Reference picture.
The principle of normalized crosscorrelation matching algorithm is: the size of image to be detected S is M1×M2, the size of template image T
For N1×N2, general M > > N, i.e. M1Greater than N1、M2Greater than N2, template image T on image to be detected S from left to right, on to
Under search for pixel-by-pixel, the subgraph that search window is covered is denoted as Si,j, wherein (i, j) is the top left corner apex of subgraph to be detected
Scheme the coordinate in S.The gray scale related coefficient between each subgraph and Prototype drawing is calculated by correlation function R (i, j), coefficient is maximum
Subgraph coordinate be matching position.
Step 3: being detected and matched with SIFT feature.
Step3.1: establishing scale space, detects key point:
The scale space function L (x, y, σ) of one sub-picture I (x, y) can be defined as making image the Gauss of variable dimension
Convolution:
L (x, y, σ)=G (x, y, σ) * I (x, y)
Wherein G (x, y, σ) is dimensional Gaussian kernel function, is indicated are as follows:
In order to effectively detect stable key point in scale space, need by difference of Gaussian DoG
(Difference of Gaussians) constructs difference of Gaussian pyramid, that is, utilizes the Gaussian difference pyrene and image of different scale
Carry out convolution:
D (x, y, σ)=(G (x, y, k σ)-G (x, y, σ)) * (x, y)=L (x, y, k σ)-L (x, y, σ)
Key point is made of the Local Extremum in the space DoG, and tentatively detecting for key point is by each in same group
Compare completion between DoG adjacent two layers image.Find DoG Function Extreme Value point, the phase that each pixel will be all with it
Adjoint point compares, and sees whether it is bigger than the consecutive points of its image area and scale domain or small.Intermediate test point and its same scale
8 consecutive points and corresponding 9 × 2 points of neighbouring scale totally 26 points compare, to ensure in scale space and X-Y scheme
Image space all detects extreme point.
There is stronger skirt response since DoG value is more sensitive to noise and edge, and in image border.It is detected to above-mentioned
Extreme point eliminate edge effect, and remove low-contrast point, obtain more accurate extreme point namely key point.
Step3.2: the distribution of key point direction generates key point description:
The gradient for seeking each extreme point assigns direction according to the peak value of histogram of gradients for extreme point.
The gradient of pixel indicates:
Gradient magnitude:
Gradient direction:
To image-region piecemeal around key point, calculation block inside gradient histogram generates unique vector, this
Vector is that one kind of the regional image information is abstract, has uniqueness.
Step3.3: key point matching:
Subclass is described to key point is established with reference to figure and floating figure, calculates similarity next of description of two set
With key point.In the case where characteristic point is few, the matching of key point directlys adopt the method for exhaustion to complete.Matching result such as Fig. 5 institute
Show.
With reference to key point description in figure:
Ri=(ri1,ri2,L,ri128,)
Key point description in floating figure:
Si=(si1,si2,L,si128,)
Any two describe sub- similarity measurement:
Key point description matched, d (Ri,Si) need to meet:
Critical value Threshold is 0.6-0.75；
Step 4:MLESAC algorithm rejects Mismatching point.
MLESAC algorithm is to introduce Mind on statistics on the basis of random sampling consistency (RANSAC) algorithm, pass through
The innovatory algorithm that likelihood score is assessed the model that generates to random sampling.
MLESAC algorithm indicates that error is distributed using mixture probabilistic model:
Wherein e is evaluated error, and σ is Gaussian Profile variance, and γ is probabilistic model weighted value, and v is equally distributed search
Window constant.
It can be seen that interior point distribution using Gaussian Profile from the probabilistic model of above formula, and exterior point (outliers) is adopted
Then it is uniformly distributed.And the method that MLESAC algorithm still uses random sampling, such a solution maximal possibility estimation
Problem, which can be converted to, solves cost function minimum problems.Cost function is shown below.MLESAC algorithm removes error hiding
Result it is as shown in Figure 6.
Step 5: transformation parameter estimation.
If some coordinates are (x0,y0), then coordinate (the x after affine transformation1,y1) can be expressed in matrix as:
[x1 y11]=[x0 y0 1]×T
Wherein T is affine matrix:
Wherein t11,t12,t21,t22It is rotation and zoom scale parameter, and tx,tyIt is translation parameters.Simply by the presence of 3 couples of spies
Levy point, so that it may which simultaneous equations solve affine parameter；It is asked when there are 4 pairs or more numbers to characteristic point using least square method
Solve more accurate affine parameter.Registration result is as shown in Figure 7.
Further, if the affine parameter that step 5 solves after executing once does not reach requirement accurately, by step 5
Execute it is primary after obtained registration result as floating figure repeat the above steps 3,4,5, after the wheel of iteration 3 to 5, obtain high-precision
Affine parameter.
The variation diagram of match point logarithm after Fig. 8 shows iteration repeatedly, with first group of data instance in Fig. 7.Fig. 9 is shown
Iteration repeatedly after root-mean-square error variation diagram, with first group of data instance in Fig. 7.Multiple the 1 meter new vacuum infrared sun is hoped
Registration result such as Figure 10 institute that the Regional High Resolution observed image picture and the full-time face observed image of SDO/HMI of remote mirror observation carry out
Show.
In conjunction with attached drawing, the embodiment of the present invention is explained in detail above, but the present invention is not limited to above-mentioned
Embodiment within the knowledge of a person skilled in the art can also be before not departing from present inventive concept
Put that various changes can be made.
Claims (3)
1. a kind of sun image method for registering based on normalized crosscorrelation and SIFT, characterized by the following steps: first
First one group is carried out respectively from the full-time face observed image of SDO and the sun local high resolution observed image of NVST down-sampled
Pretreatment；Then normalized crosscorrelation matching algorithm is used, visual field inconsistence problems in region subject to registration is solved, calculates full heliographic chart
As the best match position with Regional High Resolution image, subgraph is intercepted as reference picture to be registered using the position；Using base
Feature detection is carried out to image subject to registration in the feature detective operators (SIFT) of Scale invariant, obtains feature point set；It uses
MLESAC algorithm eliminates error hiding characteristic point pair；The transformation parameter between image finally is solved using least square method, is registrated
As a result.
2. the sun image method for registering according to claim 1 based on normalized crosscorrelation and SIFT, it is characterised in that:
Specific step is as follows for the method:
Step 1: two width sun observation images of input, wherein a width is the full-time face observed image from SDO, as reference picture
I1, another width is the sun local high resolution observed image from NVST, as floating image I2, to two images respectively into
Row pretreatment, down-sampled to two images progress first, the full-time figure of SDO reduces 0.5 times, NVST local high resolution image down
0.1 times, using rayleigh distributed, enhancing coefficient is 0.2 couple of two images I1、I2Carry out limitation contrast histogram equalization enhancing；
Step 2: using normalized crosscorrelation matching algorithm, solve visual field inconsistence problems in region subject to registration, calculate full heliographic chart
As the best match position with Regional High Resolution image, subgraph is intercepted as reference picture to be registered using the position；
The principle of normalized crosscorrelation matching algorithm is: the size of image to be detected S is M1×M2, the size of template image T is N1
×N2, M1Greater than N1、M2Greater than N2, template image T searches for pixel-by-pixel from left to right, from top to bottom on image to be detected S, searches for
The subgraph that window is covered is denoted as Si,j, wherein (i, j) is coordinate of the top left corner apex of subgraph in mapping S to be checked, pass through phase
The gray scale related coefficient between each subgraph of function R (i, j) calculating and Prototype drawing is closed, the maximum subgraph coordinate of coefficient is to match
Position:
Step 3: it is detected and is matched with SIFT feature:
Step3.1: establishing scale space, detects key point:
The scale space function L (x, y, σ) of one sub-picture I (x, y) can be defined as making image the Gaussian convolution of variable dimension:
L (x, y, σ)=G (x, y, σ) * I (x, y)
Wherein G (x, y, σ) is dimensional Gaussian kernel function, is indicated are as follows:
By difference of Gaussian DoG construct difference of Gaussian pyramid, i.e., using different scale Gaussian difference pyrene and image into
Row convolution:
D (x, y, σ)=(G (x, y, k σ)-G (x, y, σ)) * I (x, y)=L (x, y, k σ)-L (x, y, σ)
Key point is made of the Local Extremum in the space DoG, and tentatively detecting for key point is by DoG phase each in same group
Compare completion between adjacent two tomographic images, in order to find DoG Function Extreme Value point, each pixel is all adjacent with it
Point compares, and sees whether it is bigger than the consecutive points of its image area and scale domain or small, and intermediate test point is with it with the 8 of scale
Totally 26 points compare for a consecutive points and corresponding 9 × 2 points of neighbouring scale, empty in scale space and two dimensional image to ensure
Between all detect extreme point；
There is stronger skirt response since DoG value is more sensitive to noise and edge, and in image border, to the above-mentioned pole detected
Value point eliminates edge effect, and removes low-contrast point, so that it may obtain more accurate extreme point namely key point；
Step3.2: the distribution of key point direction generates key point description:
The gradient for seeking each extreme point assigns direction according to the peak value of histogram of gradients for extreme point；
The gradient of pixel indicates:
Gradient magnitude:
Gradient direction:
To image-region piecemeal around key point, calculation block inside gradient histogram generates unique vector, this vector
Be the regional image information one kind it is abstract, there is uniqueness；
Step3.3: key point matching:
Subclass is described to key point is established with reference to figure and floating figure, the similarity for calculating description of two set is closed to match
Key point, in the case where characteristic point is few, the matching of key point directlys adopt the method for exhaustion to complete；
With reference to key point description in figure:
Ri=(ri1,ri2,L,ri128,)
Key point description in floating figure:
Si=(si1, si2, L, si128)
Any two describe sub- similarity measurement:
Key point description matched, d (Ri, Si) need to meet:
Critical value Threshold is 0.6-0.75；
Step 4:MLESAC algorithm rejects Mismatching point:
MLESAC algorithm indicates that error is distributed using mixture probabilistic model:
Wherein e is evaluated error, and σ is Gaussian Profile variance, and γ is probabilistic model weighted value, and v is equally distributed search window
Constant；
It can be seen that interior point distribution using Gaussian Profile from the probabilistic model of above formula, and exterior point outliers is used then
It is to be uniformly distributed, and the method that MLESAC algorithm still uses random sampling, such a solution maximal possibility estimation problem can
Cost function minimum problems are solved to be converted to, cost function is shown below:
Step 5: transformation parameter estimation:
If some coordinates are (x0, y0), then coordinate (the x after affine transformation1, y1) can be expressed in matrix as:
[x1 y11]=[x0 y0 1]×T
Wherein T is affine matrix:
Wherein t11,t12,t21,t22It is rotation and zoom scale parameter, and tx,tyIt is translation parameters, simply by the presence of 3 pairs of characteristic points,
Can simultaneous equations solve affine parameter；It is solved when there are 4 pairs or more numbers to characteristic point using least square method
Accurate affine parameter.
3. the sun image method for registering according to claim 2 based on normalized crosscorrelation and SIFT, it is characterised in that:
If the affine parameter that step 5 solves after executing once does not reach requirement accurately, what is obtained after step 5 is executed once matches
Quasi- result repeats the above steps 3,4,5 as floating figure, after iteration 3 to 5 is taken turns, obtains high-precision affine parameter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811558754.3A CN109785371A (en) | 2018-12-19 | 2018-12-19 | A kind of sun image method for registering based on normalized crosscorrelation and SIFT |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811558754.3A CN109785371A (en) | 2018-12-19 | 2018-12-19 | A kind of sun image method for registering based on normalized crosscorrelation and SIFT |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109785371A true CN109785371A (en) | 2019-05-21 |
Family
ID=66497306
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811558754.3A Pending CN109785371A (en) | 2018-12-19 | 2018-12-19 | A kind of sun image method for registering based on normalized crosscorrelation and SIFT |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109785371A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110782394A (en) * | 2019-10-21 | 2020-02-11 | 中国人民解放军63861部队 | Panoramic video rapid splicing method and system |
| CN111028276A (en) * | 2019-12-09 | 2020-04-17 | Oppo广东移动通信有限公司 | Image alignment method and device, storage medium and electronic equipment |
| CN111179185A (en) * | 2019-12-04 | 2020-05-19 | 中国资源卫星应用中心 | Remote sensing image color correction method and system based on cloud pickling film and MSR |
| CN111311525A (en) * | 2019-11-20 | 2020-06-19 | 重庆邮电大学 | Image gradient field double-interval equalization algorithm based on histogram probability correction |
| CN111583320A (en) * | 2020-03-17 | 2020-08-25 | 哈尔滨医科大学 | Breast cancer ultrasonic image typing method and system fusing deep convolutional network and image omics characteristics and storage medium |
| CN112307981A (en) * | 2020-10-29 | 2021-02-02 | 西北工业大学 | Feature information transmission and cooperative tracking method in space rolling non-cooperative target observation process |
| CN112488240A (en) * | 2020-12-16 | 2021-03-12 | 广东省科学院智能制造研究所 | Image frequency domain difference material matching method based on cross correlation |
| CN112488958A (en) * | 2020-12-15 | 2021-03-12 | 西安交通大学 | Image contrast enhancement method based on scale space |
| CN112819095A (en) * | 2021-02-26 | 2021-05-18 | 吉林大学 | Feature point matching method and device, intelligent terminal and computer readable storage medium |
| CN113191935A (en) * | 2021-04-30 | 2021-07-30 | 华中科技大学 | Reconfigurable hardware acceleration method and system for Gaussian pyramid construction |
| CN113379808A (en) * | 2021-06-21 | 2021-09-10 | 昆明理工大学 | Method for registration of multiband solar images |
| CN115862259A (en) * | 2023-02-24 | 2023-03-28 | 深圳市特安电子有限公司 | Fire alarm early warning system based on temperature monitoring |
| CN116128734A (en) * | 2023-04-17 | 2023-05-16 | 湖南大学 | Image stitching method, device, equipment and medium based on deep learning |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020967A (en) * | 2012-12-07 | 2013-04-03 | 青岛经纬蓝图信息技术有限公司 | Unmanned aerial vehicle aerial image accurate matching method based on island edge characteristics |
| CN103729654A (en) * | 2014-01-22 | 2014-04-16 | 青岛新比特电子科技有限公司 | Image matching retrieval system on account of improving Scale Invariant Feature Transform (SIFT) algorithm |
| CN103839265A (en) * | 2014-02-26 | 2014-06-04 | 西安电子科技大学 | SAR image registration method based on SIFT and normalized mutual information |
| CN105469110A (en) * | 2015-11-19 | 2016-04-06 | 武汉大学 | Non-rigid transformation image characteristic matching method based on local linear transfer and system |
| CN106408600A (en) * | 2016-09-08 | 2017-02-15 | 昆明理工大学 | Image registration method applied to solar high-resolution image |
-
2018
- 2018-12-19 CN CN201811558754.3A patent/CN109785371A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020967A (en) * | 2012-12-07 | 2013-04-03 | 青岛经纬蓝图信息技术有限公司 | Unmanned aerial vehicle aerial image accurate matching method based on island edge characteristics |
| CN103729654A (en) * | 2014-01-22 | 2014-04-16 | 青岛新比特电子科技有限公司 | Image matching retrieval system on account of improving Scale Invariant Feature Transform (SIFT) algorithm |
| CN103839265A (en) * | 2014-02-26 | 2014-06-04 | 西安电子科技大学 | SAR image registration method based on SIFT and normalized mutual information |
| CN105469110A (en) * | 2015-11-19 | 2016-04-06 | 武汉大学 | Non-rigid transformation image characteristic matching method based on local linear transfer and system |
| CN106408600A (en) * | 2016-09-08 | 2017-02-15 | 昆明理工大学 | Image registration method applied to solar high-resolution image |
Non-Patent Citations (1)
| Title |
|---|
| 焦继超 等: ""一种基于非降采样 Contourlet 变换和 MLESAC的星空图像配准算法"", 《系统工程与电子技术》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110782394A (en) * | 2019-10-21 | 2020-02-11 | 中国人民解放军63861部队 | Panoramic video rapid splicing method and system |
| CN111311525A (en) * | 2019-11-20 | 2020-06-19 | 重庆邮电大学 | Image gradient field double-interval equalization algorithm based on histogram probability correction |
| CN111179185B (en) * | 2019-12-04 | 2023-07-18 | 中国四维测绘技术有限公司 | Remote sensing image color correction method and system based on cloud pickling film and MSR |
| CN111179185A (en) * | 2019-12-04 | 2020-05-19 | 中国资源卫星应用中心 | Remote sensing image color correction method and system based on cloud pickling film and MSR |
| CN111028276A (en) * | 2019-12-09 | 2020-04-17 | Oppo广东移动通信有限公司 | Image alignment method and device, storage medium and electronic equipment |
| CN111583320A (en) * | 2020-03-17 | 2020-08-25 | 哈尔滨医科大学 | Breast cancer ultrasonic image typing method and system fusing deep convolutional network and image omics characteristics and storage medium |
| CN112307981A (en) * | 2020-10-29 | 2021-02-02 | 西北工业大学 | Feature information transmission and cooperative tracking method in space rolling non-cooperative target observation process |
| CN112488958A (en) * | 2020-12-15 | 2021-03-12 | 西安交通大学 | Image contrast enhancement method based on scale space |
| CN112488240A (en) * | 2020-12-16 | 2021-03-12 | 广东省科学院智能制造研究所 | Image frequency domain difference material matching method based on cross correlation |
| CN112819095A (en) * | 2021-02-26 | 2021-05-18 | 吉林大学 | Feature point matching method and device, intelligent terminal and computer readable storage medium |
| CN113191935A (en) * | 2021-04-30 | 2021-07-30 | 华中科技大学 | Reconfigurable hardware acceleration method and system for Gaussian pyramid construction |
| CN113379808A (en) * | 2021-06-21 | 2021-09-10 | 昆明理工大学 | Method for registration of multiband solar images |
| CN113379808B (en) * | 2021-06-21 | 2022-08-12 | 昆明理工大学 | Method for registration of multiband solar images |
| CN115862259A (en) * | 2023-02-24 | 2023-03-28 | 深圳市特安电子有限公司 | Fire alarm early warning system based on temperature monitoring |
| CN116128734A (en) * | 2023-04-17 | 2023-05-16 | 湖南大学 | Image stitching method, device, equipment and medium based on deep learning |
| CN116128734B (en) * | 2023-04-17 | 2023-06-23 | 湖南大学 | Image stitching method, device, equipment and medium based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109785371A (en) | A kind of sun image method for registering based on normalized crosscorrelation and SIFT | |
| Li et al. | LNIFT: Locally normalized image for rotation invariant multimodal feature matching | |
| CN103914847B (en) | Based on phase equalization and the SAR image registration method of SIFT | |
| CN104574347B (en) | Satellite in orbit image geometry positioning accuracy evaluation method based on multi- source Remote Sensing Data data | |
| CN104134200B (en) | Mobile scene image splicing method based on improved weighted fusion | |
| CN106651904B (en) | A kind of more extraterrestrial target method for capturing and tracing of width size range | |
| CN106529538A (en) | Method and device for positioning aircraft | |
| CN105957054B (en) | A kind of image change detection method | |
| CN111476251A (en) | Remote sensing image matching method and device | |
| CN112288008A (en) | Mosaic multispectral image disguised target detection method based on deep learning | |
| CN106530271A (en) | Infrared image significance detection method | |
| CN113538347B (en) | Image detection method and system based on efficient bidirectional path aggregation attention network | |
| CN114066955A (en) | Registration method for registering infrared light image to visible light image | |
| CN109117851A (en) | A kind of video image matching process based on lattice statistical constraint | |
| CN108680177A (en) | Synchronous superposition method and device based on rodent models | |
| CN110793529B (en) | Quick matching star map identification method | |
| CN108320310A (en) | Extraterrestrial target 3 d pose method of estimation based on image sequence | |
| CN105631849B (en) | The change detecting method and device of target polygon | |
| CN110060199A (en) | A kind of quick joining method of plant image based on colour and depth information | |
| CN108921899A (en) | A kind of indoor vision positioning method solving fundamental matrix based on pixel threshold | |
| CN109978982B (en) | Point cloud rapid coloring method based on oblique image | |
| CN111126508A (en) | Hopc-based improved heterogeneous image matching method | |
| CN105225232A (en) | A kind of colour of view-based access control model attention mechanism warship cooperative target detection method | |
| CN109190506A (en) | It is a kind of based on core is sparse and the EO-1 hyperion object detection method of space constraint | |
| CN111899284B (en) | Planar target tracking method based on parameterized ESM network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190521 |
|
| RJ01 | Rejection of invention patent application after publication |