CN108961164A - Image registration joining method and device and computer equipment based on geometrical invariants - Google Patents
Image registration joining method and device and computer equipment based on geometrical invariants Download PDFInfo
- Publication number
- CN108961164A CN108961164A CN201810730583.1A CN201810730583A CN108961164A CN 108961164 A CN108961164 A CN 108961164A CN 201810730583 A CN201810730583 A CN 201810730583A CN 108961164 A CN108961164 A CN 108961164A
- Authority
- CN
- China
- Prior art keywords
- point
- registration
- feature points
- group
- feature
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- 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
- G06T3/40—Scaling the whole image or part thereof
- G06T3/4038—Scaling the whole image or part thereof for image mosaicing, i.e. plane images composed of plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/32—Indexing scheme for image data processing or generation, in general involving image mosaicing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The image registration joining method and device and computer equipment that the invention discloses a kind of based on geometrical invariants.This method comprises: carrying out binary conversion treatment to reference star chart and star chart subject to registration；Calculate the characteristic point coordinate for referring to star chart and star chart subject to registration；Construct the geometrical invariants of all characteristic points；Matching cost function is constructed according to geometrical invariants to the asterism of reference star chart and star chart subject to registration；Cost function is calculated, the characteristic point of reference star chart and star chart subject to registration is matched；N group matching double points are randomly selected, calculate registration parameter according to rigid body translation model；Remaining matching double points registration parameter is verified, error hiding is removed.By means of the invention it is possible to improve registration accuracy, meet the needs of subsequent image co-registration or image mosaic.
Description
Technical field
The present invention relates to star image processing technical fields, more particularly, to a kind of image registration based on geometrical invariants
Joining method and device and computer equipment.
Background technique
Image registration is under the visual angle found in different time points, different or by different sensors shooting about same field
Spatial transform relation between the two images or multiple image of scape, and one or more therein are matched and are superimposed
Process.Image registration can simply be interpreted as the process of a searching space reflection, rearrange the pixel point of piece image
It sets, and keeps consistency spatially with the corresponding pixel points of another piece image.Image registration is remote sensing image processing, target knowledge
Not, one of the key technology in numerous art of image analysis such as image co-registration, image mosaic, image reconstruction, robot vision is
Research hotspot in field of image processing.
Image registration is broadly divided into based on region and based on the image registration of feature, and the method used when for matching is not
It is same further to be divided again.Method for registering images based on region is also known as template matching method, it is without the concern for figure
The structural information of picture matches the grayscale information of image using the relevant method in region.But it becomes complicated image
Change nearly unavailable, while the computation complexity of algorithm is higher.Method for registering images based on region has cross-correlation method (Cross
Correlation, CC), sequential detection method (Similarity Detection Algorithm, SSDA) is based in a frequency domain
The Fourier-Mellin transformation phase correlation method of Fourier transform, mutual information method (Mutual Information, MI),
Levenberg-Marquardt optimization method；Method for registering images based on feature be not by all image-regions but
The matching between image is carried out by characteristic information representative on image, to achieve the purpose that image registration.Due to
The method for registering only considers the matching of characteristic information, and stronger to the anti-interference ability of noise, deformation, thus has smaller
Calculation amount and higher robustness.Method for registering images based on feature has the arest neighbors iteration point using spatial relationship
(Iterative Cloest Point, ICP) method, the method using constant description, consistency label (Consistent
Labeling Problem, CLP) relaxation method, pyramid and wavelet method etc..
The method of image registration at present mainly has image registration joining method based on Sift characteristic point and based in shape
Image registration joining method hereafter.Since the description vectors difference of the Sift characteristic point near the same pixel is larger, lead
The characteristic point between two images is caused not match correctly；And since the time for exposure is different from ISO (sensitivity), lead to two width
Figure extracts obtained characteristic point quantity and differs larger, incoherent even the shape histogram of the same point also differs more
Point is but possible to similar, thus registration is caused to fail.
It is therefore proposed that a kind of new image registration joining method, the phenomenon that reducing characteristic point mismatch, registration accuracy is improved,
As urgent problem to be solved.
Summary of the invention
In view of this, the image registration joining method and device that the present invention provides a kind of based on geometrical invariants and calculating
Machine equipment and computer readable storage medium make full use of in star chart existing geometrical invariants information between asterism, in star chart
Existing characteristic point is established Feature Descriptor and is matched, and it is low to solve characteristic point mismatch and registration accuracy in the prior art
Technical problem.
In order to solve the above-mentioned technical problem, according to an aspect of the invention, there is provided it is a kind of based on geometrical invariants
Image registration joining method.
Method includes the following steps:
Step S1: carrying out binary conversion treatment to reference star chart and star chart subject to registration respectively, obtains with reference to M star in star chart
N number of asterism in point and star chart subject to registration, wherein M and N is the natural number more than or equal to 4；
Step S2: the corresponding characteristic point of each asterism is calculated；
Step S3: the coordinate of each characteristic point is calculated, wherein with reference to the seat of the corresponding characteristic point of M asterism in star chart
Mark constitutes fixed reference feature point set, and the coordinate of the corresponding characteristic point of N number of asterism in star chart subject to registration constitutes feature point set subject to registration；
Step S4: it is concentrated in fixed reference feature point, calculates each characteristic point to the distance metric of remaining characteristic point, referred to
The geometrical invariants description vectors of each characteristic point in feature point set calculate each characteristic point and arrive in feature point set subject to registration
The distance metric of remaining characteristic point obtains the geometrical invariants description vectors of each characteristic point in feature point set subject to registration, wherein
It includes M-1 distance metric, feature point set subject to registration that fixed reference feature point, which concentrates the geometrical invariants description vectors of each characteristic point,
In the geometrical invariants description vectors of each characteristic point include N-1 distance metric；
Step S5: fixed reference feature point is concentrated to the geometrical invariants description vectors and feature point set subject to registration of a characteristic point
In the geometrical invariants description vectors of characteristic point input preset matching cost function, obtain the matching of two characteristic points
Value；
Step S6: judging whether matching value is greater than preset matching threshold value, when matching value is greater than preset matching threshold value, matching
Being worth corresponding two characteristic points is one group of Feature Points Matching point pair, obtains S group Feature Points Matching point pair；
Step S7: N1 group Feature Points Matching point pair is extracted in S group Feature Points Matching point pair, to each group of Feature Points Matching
Point obtains each group of Feature Points Matching point to corresponding registration parameter to calculating respectively according to rigid body translation model；
Step S8: for each group of Feature Points Matching point to corresponding registration parameter, it is utilized respectively N2 group Feature Points Matching
Whether point is to registration parameter is verified, to determine this group of Feature Points Matching point to correct, wherein N1+N2=S, and N2 group
Feature Points Matching point is to the Feature Points Matching point pair for the point centering of S group Feature Points Matching in addition to N1 group Feature Points Matching point pair；
Step S9: by the incorrect Feature Points Matching point of N1 group Feature Points Matching point centering to removal, it is correct to obtain X group
Feature Points Matching point pair；
Step S10: it is calculated using the following equation iteration threshold r:
Wherein, p is the correct point of expectation to probability, SnIt is the Feature Points Matching point that is obtained after nth iteration to sum,Sn=Sn-1+ X-X', S0=0, X' are the Feature Points Matching point that obtains after (n-1)th iteration to the corresponding feature of sum
In point matching double points, the number of Feature Points Matching point pair identical with the correct Feature Points Matching point centering of X group；
Step S11: judging whether iteration threshold r is greater than or equal to 1, if iteration threshold r is greater than or equal to 1, and will repeatedly
For frequency n plus 1, return step S7, wherein the initial value of n is 0, if iteration threshold r less than 1, terminates iteration, nth iteration
The Feature Points Matching point obtained afterwards is to the corresponding Feature Points Matching point of sum to for the match point with reference to star chart and star chart subject to registration.
Further, step S1 includes: that reference star chart and star chart subject to registration are respectively converted into grey level histogram；To reference
The corresponding grey level histogram of star chart is calculated using least square method, obtains the first optimal threshold, wherein the first optimal threshold makes to join
The variance for examining the asterism region and background area in star chart is maximum；To the corresponding grey level histogram of star chart subject to registration using minimum two
Multiplication calculates, and obtains the second optimal threshold, wherein the second optimal threshold makes asterism region and background area in star chart subject to registration
Variance it is maximum；Binary conversion treatment is carried out to reference star chart according to the first optimal threshold, is obtained with reference to M asterism in star chart；
Binary conversion treatment is carried out to star chart subject to registration according to the second optimal threshold, obtains N number of asterism in star chart subject to registration.
Further, step S2 includes: that the mass center of asterism on the image after calculating binarization segmentation obtains characteristic point.
Further, in step s 4, it is calculated using the following equation and is concentrated in the fixed reference feature point, each characteristic point arrives
The distance metric of remaining characteristic point:
lpi(k)=| | pi-pk||2, k=1,2..., M
Wherein, k ≠ i,For the fixed reference feature point set, piAnd pkIt is the fixed reference feature point set
The coordinate of middle characteristic point, lpiIt (k) is piTo pkDistance metric；
It is calculated using the following equation in the feature point set subject to registration, the distance degree of each characteristic point to remaining characteristic point
Amount:
lqj(l)=| | qj-ql||2, l=1,2 ..., N
Wherein, and l ≠ j,For the feature point set subject to registration, qjAnd qlIt is the spy subject to registration
Sign point concentrates the coordinate of characteristic point, lqjIt (l) is qjTo qlDistance metric.
Further, the matching cost function of construction is as follows:
Wherein, δ is the standard deviation of Gaussian function, EijFor piWith qjMatching value.
Further, step S7 includes: to each group of Feature Points Matching point to calculating rotation, translation, change of scale and throwing
The parameter of shadow transformation, obtains each group of Feature Points Matching point to corresponding registration parameter.
Further, step S8 includes: to be verified using N2 group Feature Points Matching point to registration parameter, if error is small
In default error threshold, then it is assumed that this group of Feature Points Matching point is correct.
In order to solve the above-mentioned technical problem, according to the second aspect of the invention, provide a kind of based on geometrical invariants
Image registration splicing apparatus.
The device comprises the following modules:
Binarization block is obtained for carrying out binary conversion treatment respectively to reference star chart and star chart subject to registration with reference to star chart
In M asterism and star chart subject to registration in N number of asterism, wherein M and N is the natural number more than or equal to 4；
Characteristic point computing module, for calculating the corresponding characteristic point of each asterism；
Coordinate calculation module, for calculating the coordinate of each characteristic point, wherein corresponding with reference to M asterism in star chart
The coordinate of characteristic point constitutes fixed reference feature point set, and the coordinate of the corresponding characteristic point of N number of asterism in star chart subject to registration is constituted wait match
Quasi- feature point set；
Description vectors computing module, for being concentrated in fixed reference feature point, calculate each characteristic point to remaining characteristic point away from
From measurement, the geometrical invariants description vectors that fixed reference feature point concentrates each characteristic point are obtained, in feature point set subject to registration, meter
Each characteristic point is calculated to the distance metric of remaining characteristic point, obtains the geometrical invariants of each characteristic point in feature point set subject to registration
Description vectors, wherein it includes M-1 distance metric that fixed reference feature point, which concentrates the geometrical invariants description vectors of each characteristic point,
The geometrical invariants description vectors of each characteristic point include N-1 distance metric in feature point set subject to registration；
Matching value computing module, for fixed reference feature point is concentrated characteristic point geometrical invariants description vectors and to
Registration features point concentrates the geometrical invariants description vectors an of characteristic point to input preset matching cost function, obtains two spies
Levy the matching value of point；
Matching double points determining module is preset for judging whether matching value is greater than preset matching threshold value when matching value is greater than
When matching threshold, corresponding two characteristic points of matching value are one group of Feature Points Matching point pair, obtain S group Feature Points Matching point pair；
Registration parameter computing module, for extracting N1 group Feature Points Matching point pair in S group Feature Points Matching point pair, to every
One group of Feature Points Matching point obtains each group of Feature Points Matching point to corresponding to calculating respectively according to rigid body translation model
Registration parameter；
Authentication module, for, to corresponding registration parameter, being utilized respectively N2 group feature for each group of Feature Points Matching point
Whether point matching double points verify registration parameter, to determine this group of Feature Points Matching point to correct, wherein N1+N2=S,
And N2 group Feature Points Matching point is to the Feature Points Matching for the point centering of S group Feature Points Matching in addition to N1 group Feature Points Matching point pair
Point pair；
Error hiding removes module, for by the incorrect Feature Points Matching point of N1 group Feature Points Matching point centering to removal,
Obtain the correct Feature Points Matching point pair of X group；
Iteration threshold computing module, for being calculated using the following equation iteration threshold r:
Wherein, p is the correct point of expectation to probability, SnIt is the Feature Points Matching point that is obtained after nth iteration to sum,Sn=Sn-1+ X-X', S0=0, X' are the Feature Points Matching point that obtains after (n-1)th iteration to the corresponding feature of sum
In point matching double points, the number of Feature Points Matching point pair identical with the correct Feature Points Matching point centering of X group；
Iteration control module, for judging whether iteration threshold r is greater than or equal to 1, if iteration threshold r is greater than or equal to 1,
Then and by the number of iterations n add 1, and calls registration parameter computing module, authentication module, error hiding removal module and iteration threshold meter
It calculates module to be iterated, wherein the initial value of n is 0, if iteration threshold r less than 1, terminates iteration, is obtained after nth iteration
Feature Points Matching point to the corresponding Feature Points Matching point of sum to for the match point with reference to star chart and star chart subject to registration.
In order to solve the above-mentioned technical problem, according to the third aspect of the present invention, a kind of computer equipment is provided.
The computer equipment include memory, processor and storage on a memory and the meter that can run on a processor
Calculation machine program when the processor executes the program, realizes that any one image based on geometrical invariants provided by the invention is matched
The step of quasi- method.
In order to solve the above-mentioned technical problem, according to the fourth aspect of the present invention, a kind of computer-readable storage is provided
Medium.
It is stored with computer program on the computer readable storage medium, the present invention is realized when which is executed by processor
The step of any one method for registering images based on geometrical invariants provided.
Compared with prior art, it is provided by the invention by the image registration joining method and device of geometrical invariants and based on
Calculate machine equipment and computer readable storage medium, at least realize it is following the utility model has the advantages that
First, the image registration joining method of the present invention based on geometrical invariants is introduced when characteristic point is chosen
The method of geometrical invariants, this method have the characteristics that adaptivity, can be good at overcoming in star chart lack texture information and
There are the difficulties of ambient noise, accurately choose suitable characteristic point and are matched, achieve preferable effect.
Second, the image registration joining method of the present invention based on geometrical invariants, in reference star chart and star subject to registration
Caused by the characteristic point quantity difference of figure is larger in the different situation of feature vector dimension, construct one based on it is several why not
The matching cost function of variable obtains matching characteristic point pair, compared to Sift method and based on the method for Shape context, energy
More characteristic points pair are accessed, the requirement of image registration is better met.
Third, the image registration joining method of the present invention based on geometrical invariants are removed using by successive ignition
The characteristic point pair for falling mistake, has reached sub-pixed mapping precision, has met the needs of subsequent image co-registration or image mosaic.
Certainly, implementing any of the products of the present invention specific needs while must not reach all the above technical effect.
By referring to the drawings to the detailed description of exemplary embodiment of the present invention, other feature of the invention and its
Advantage will become apparent.
Detailed description of the invention
It is combined in the description and the attached drawing for constituting part of specification shows the embodiment of the present invention, and even
With its explanation together principle for explaining the present invention.
Fig. 1 is the step process for the image registration joining method based on geometrical invariants that the embodiment of the present invention one provides
Figure；
Fig. 2 is the block diagram of the image registration splicing apparatus provided by Embodiment 2 of the present invention based on geometrical invariants；
Fig. 3 is the hardware structure diagram for the computer equipment that the embodiment of the present invention three provides.
Specific embodiment
Carry out the various exemplary embodiments of detailed description of the present invention now with reference to attached drawing.It should also be noted that unless in addition having
Body explanation, the unlimited system of component and the positioned opposite of step, numerical expression and the numerical value otherwise illustrated in these embodiments is originally
The range of invention.
Be to the description only actually of at least one exemplary embodiment below it is illustrative, never as to the present invention
And its application or any restrictions used.
Technology, method and apparatus known to person of ordinary skill in the relevant may be not discussed in detail, but suitable
In the case of, the technology, method and apparatus should be considered as part of specification.
It is shown here and discuss all examples in, any occurrence should be construed as merely illustratively, without
It is as limitation.Therefore, other examples of exemplary embodiment can have different values.
It should also be noted that similar label and letter indicate similar terms in following attached drawing, therefore, once a certain Xiang Yi
It is defined in a attached drawing, then in subsequent attached drawing does not need that it is further discussed.
Embodiment one
The embodiment of the present invention one provides a kind of image registration joining method based on geometrical invariants, the Registration and connection side
Method can determine star chart subject to registration and the Feature Points Matching point pair with reference to star chart, to complete star chart subject to registration and with reference to star chart
Registration, Fig. 1 are the step flow chart for the image registration joining method based on geometrical invariants that the embodiment of the present invention one provides, tool
Body, as shown in Figure 1, this method includes the following steps, namely S1 to step S11.
Step S1: carrying out binary conversion treatment to reference star chart and star chart subject to registration respectively, obtains with reference to M star in star chart
N number of asterism in point and star chart subject to registration.
Wherein, M and N is the natural number more than or equal to 4.On reference star chart and star chart subject to registration, asterism location
The pixel in domain is different with the gray scale of the pixel of the background region in addition to asterism, thus, to reference star chart and star subject to registration
After figure carries out binary conversion treatment respectively, asterism and background can be distinguished, existing skill can be used to obtain asterism, at this
Arbitrary binary processing method in art, it is preferable that further real by specific steps below when realizing step S1
It is existing:
Step S1a: grey level histogram will be respectively converted into reference to star chart and star chart subject to registration；
Step S1b: the corresponding grey level histogram of reference star chart is calculated using least square method, obtains the first best threshold
Value, wherein the first optimal threshold keeps the variance with reference to asterism region and background area in star chart maximum；
Step S1c: the corresponding grey level histogram of star chart subject to registration is calculated using least square method, obtains the second best threshold
Value, wherein the second optimal threshold keeps the variance in asterism region and background area in star chart subject to registration maximum；
Step S1d: binary conversion treatment is carried out to reference star chart according to the first optimal threshold, is obtained with reference to M in star chart
Asterism；
Step S1e: binary conversion treatment is carried out to star chart subject to registration according to the second optimal threshold, is obtained in star chart subject to registration
N number of asterism.
Step S2: the corresponding characteristic point of each asterism is calculated.
Specifically, it calculates with reference to the corresponding characteristic point of M asterism in star chart, is calculated in star chart subject to registration in this step
The corresponding characteristic point of N number of asterism.
In this step, it when calculating the corresponding characteristic point of asterism, can be broadcast TV programs by satellite by calculating the image after binarization segmentation
The mass center of point obtains characteristic point.
Step S3: the coordinate of each characteristic point is calculated.
Wherein, fixed reference feature point set is constituted with reference to the coordinate of the corresponding characteristic point of M asterism in star chart, that is, with reference to
Feature point set includes the coordinate of M characteristic point, and the coordinate of the corresponding characteristic point of N number of asterism in star chart subject to registration constitutes subject to registration
Feature point set, that is, feature point set subject to registration includes the coordinate of N number of characteristic point.
Step S4: it is concentrated in fixed reference feature point, calculates each characteristic point to the distance metric of remaining characteristic point, referred to
The geometrical invariants description vectors of each characteristic point in feature point set calculate each characteristic point and arrive in feature point set subject to registration
The distance metric of remaining characteristic point obtains the geometrical invariants description vectors of each characteristic point in feature point set subject to registration.
Wherein, if fixed reference feature point set isIt altogether include the coordinate of M characteristic point, for each spy
Point is levied, its distance metric for arriving remaining M-1 characteristic point is calculated separately, obtains the corresponding M-1 distance metric of each characteristic point,
That is, it includes M-1 distance metric that fixed reference feature point, which concentrates the geometrical invariants description vectors of each characteristic point, in step S4
In, it is calculated using the following equation and is concentrated in fixed reference feature point, the distance metric of each characteristic point to remaining characteristic point:
lpi(k)=| | pi-pk||2, k=1,2..., M,
Wherein, k ≠ i, piAnd pkIt is the coordinate that fixed reference feature point concentrates characteristic point, lpiIt (k) is piTo pkDistance degree
Amount.
If feature point set subject to registration isIt altogether include the coordinate of N number of characteristic point, for each feature
Point calculates separately its distance metric for arriving remaining N-1 characteristic point, obtains the corresponding N-1 distance metric of each characteristic point,
The geometrical invariants description vectors of each characteristic point include N-1 distance metric in feature point set i.e. subject to registration.In step S4
In, it is calculated using the following equation in feature point set subject to registration, the distance metric of each characteristic point to remaining characteristic point:
lqj(l)=| | qj-ql||2, l=1,2..., N
Wherein, and l ≠ j, qjAnd qlIt is the coordinate of characteristic point in feature point set subject to registration, lqjIt (l) is qjTo qlDistance
Measurement.
Step S5: fixed reference feature point is concentrated to the geometrical invariants description vectors and feature point set subject to registration of a characteristic point
In the geometrical invariants description vectors of characteristic point input preset matching cost function, obtain the matching of two characteristic points
Value.
In this step, preset matching cost function, using the geometrical invariants description vectors of characteristic point, to fixed reference feature
The characteristic point in characteristic point and feature point set subject to registration that point is concentrated carries out the calculating of matching value, and fixed reference feature point is concentrated each
A matching value can be calculated in each characteristic point in a characteristic point and feature point set subject to registration, that is, fixed reference feature point
Each of collection characteristic point can correspond to calculating to N number of matching value.
Step S6: judging whether matching value is greater than preset matching threshold value, when matching value is greater than preset matching threshold value, matching
Being worth corresponding two characteristic points is one group of Feature Points Matching point pair, obtains S group Feature Points Matching point pair.
In this step, preset matching threshold value screens all matching values, filters out matching value and is greater than matching threshold
The characteristic point pair of value, each characteristic point to being named as Feature Points Matching point pair, meanwhile, the Feature Points Matching point logarithm defined
Amount is S group.
Step S7: N1 group Feature Points Matching point pair is extracted in S group Feature Points Matching point pair, to each group of Feature Points Matching
Point obtains each group of Feature Points Matching point to corresponding registration parameter to calculating respectively according to rigid body translation model.
In this step, N1 group is randomly selected out from S group Feature Points Matching point pair first, by N1 group Feature Points Matching point
In the preset rigid body translation model of each group of input calculated, N1 group registration parameter can be obtained.
Step S8: for each group of Feature Points Matching point to corresponding registration parameter, it is utilized respectively N2 group Feature Points Matching
Whether point is to registration parameter is verified, to determine this group of Feature Points Matching point to correct.
Wherein, N1+N2=S, and N2 group Feature Points Matching point removes N1 group characteristic point to for the Feature Points Matching point centering of S group
With Feature Points Matching point pair of the point to except, namely S7 through the above steps calculate obtain N1 group registration parameter and then
Using remaining N2 group Feature Points Matching point to verifying.
In a kind of verification method, when certain group registration parameter can satisfy verification condition, for example, verification condition is that N2 group is special
The Feature Points Matching point of preset percentage is all satisfied registration parameter in sign point match point, then it is assumed that the corresponding spy of this group of registration parameter
Sign point is correct characteristic point, is otherwise incorrect characteristic point.
Step S9: by the incorrect Feature Points Matching point of N1 group Feature Points Matching point centering to removal, it is correct to obtain X group
Feature Points Matching point pair.
No matter what kind of verification method is used, it is remaining by incorrect Feature Points Matching point to removal in step S
Feature Points Matching point X is defined as to quantity.
Step S10: iteration threshold r is calculated.
In this step, following formula can be used and calculate iteration threshold r:
Wherein, p is the correct point of expectation to probability, SnIt is the Feature Points Matching point that is obtained after nth iteration to sum, Sn=
Sn-1+ X-X', S0=0, X' are the Feature Points Matching point that obtains after (n-1)th iteration to the corresponding Feature Points Matching point pair of sum
In, the number of Feature Points Matching point pair identical with the correct Feature Points Matching point centering of X group.
Step S11: judging whether iteration threshold r is greater than or equal to 1, if iteration threshold r is greater than or equal to 1, and will repeatedly
For frequency n plus 1, return step S7, if iteration threshold r less than 1, terminates iteration, the Feature Points Matching obtained after nth iteration
Point is to the corresponding Feature Points Matching point of sum to for the match point with reference to star chart and star chart subject to registration.
Wherein, the initial value of n is 0, optionally, in nth iteration, first according to the S obtained after last iterationn-1
Iteration threshold e is calculated, then judges whether iteration threshold r is greater than or equal to 1, if iteration threshold r is greater than or equal to 1, then is calculated
X' specifically by the correct Feature Points Matching point pair of the X group being calculated in the nth iteration, and is calculated in last iteration
Obtained Sn-1Feature Points Matching point obtains identical Feature Points Matching point pair to being compared, these Feature Points Matching points are to i.e.
For X' group Feature Points Matching point pair, then further according to Sn=Sn-1+ X-X' calculates Sn, so as to next iteration namely (n+1)th time
Iteration threshold r is calculated in iteration.
The image registration joining method based on geometrical invariants provided using the embodiment is chosen and is matched in characteristic point
When introduce geometrical invariants and matched, enable Registration and connection method to have the characteristics that adaptivity, can be good at
Overcome and lack texture information and the difficulty there are ambient noise in star chart, accurately chooses suitable characteristic point and matched,
It is good with accuracy.
Preferably, in step S5, when calculating matching value, the matching cost function of construction is as follows:
Wherein, δ is the standard deviation of Gaussian function, EijFor piWith qjMatching value.
The image registration joining method based on geometrical invariants provided using the preferred embodiment, passes through above method structure
Matching cost function is made, it is suitable to select in the unmatched situation of characteristic set dimension of reference star chart and star chart subject to registration
Characteristic point to being registrated.
Preferably, in the step s 7, to each group of Feature Points Matching point to calculating rotation, translation, change of scale and projection
The parameter of transformation obtains each group of Feature Points Matching point to corresponding registration parameter.
The image registration joining method based on geometrical invariants provided using the preferred embodiment, geometrical invariants are in star
Figure with reference to star chart and star chart subject to registration with that will regard only rotation, translation, change of scale and projective transformation as on time, without other
Transformation simplifies calculation amount, promotes matching speed.
Preferably, in step s 8, it is verified using N2 group Feature Points Matching point to registration parameter, if error is less than
Default error threshold, then it is assumed that this group of Feature Points Matching point is correct.
Embodiment two
Second embodiment of the present invention provides a kind of image registration splicing apparatus based on geometrical invariants, Registration and connection dress
It sets and can determine star chart subject to registration and the Feature Points Matching point pair with reference to star chart, to complete star chart subject to registration and with reference to star chart
Registration, Fig. 2 is the block diagram of the image registration splicing apparatus provided by Embodiment 2 of the present invention based on geometrical invariants, specifically,
As shown in Fig. 2, the device includes: binarization block 201, characteristic point computing module 202, coordinate calculation module 203, description vectors
Computing module 204, matching value computing module 205, matching double points determining module 206, registration parameter computing module 207, verifying mould
Block 208, error hiding removal module 209, iteration threshold computing module 210 and iteration control module 211.
Wherein, binarization block 201 is joined for carrying out binary conversion treatment respectively to reference star chart and star chart subject to registration
Examine N number of asterism in the M asterism and star chart subject to registration in star chart, wherein M and N is the natural number more than or equal to 4；It is special
Sign point computing module 202 is for calculating the corresponding characteristic point of each asterism；Coordinate calculation module 203 is for calculating each characteristic point
Coordinate, wherein with reference to the corresponding characteristic point of M asterism in star chart coordinate constitute fixed reference feature point set, star chart subject to registration
In the coordinate of the corresponding characteristic point of N number of asterism constitute feature point set subject to registration；Description vectors computing module 204 is for referring to
In feature point set, each characteristic point is calculated to the distance metric of remaining characteristic point, fixed reference feature point is obtained and concentrates each characteristic point
Geometrical invariants description vectors, in feature point set subject to registration, calculate each characteristic point to remaining characteristic point distance metric,
Obtain the geometrical invariants description vectors of each characteristic point in feature point set subject to registration, wherein fixed reference feature point concentrates each spy
Sign point geometrical invariants description vectors include M-1 distance metric, in feature point set subject to registration each characteristic point it is several why not
Variable description vector includes N-1 distance metric；Matching value computing module 205 is used to fixed reference feature point concentrating a characteristic point
Geometrical invariants description vectors and feature point set subject to registration in characteristic point the input of geometrical invariants description vectors it is default
Matching cost function, obtain the matching value of two characteristic points；Matching double points determining module 206 is for judging whether matching value is big
In preset matching threshold value, when matching value is greater than preset matching threshold value, corresponding two characteristic points of matching value are one group of characteristic point
Matching double points obtain S group Feature Points Matching point pair；Registration parameter computing module 207 in S group Feature Points Matching point pair for taking out
N1 group Feature Points Matching point pair is taken, to each group of Feature Points Matching point to being calculated respectively according to rigid body translation model, is obtained
Each group of Feature Points Matching point is to corresponding registration parameter；Authentication module 208 is used for for each group of Feature Points Matching point to right
The registration parameter answered is utilized respectively N2 group Feature Points Matching point and verifies to registration parameter, to determine this group of characteristic point
With point to whether correct, wherein N1+N2=S, and N2 group Feature Points Matching point removes N1 group to for the Feature Points Matching point centering of S group
Feature Points Matching point pair except Feature Points Matching point pair；Error hiding removes module 209 and is used for N1 group Feature Points Matching point pair
In incorrect Feature Points Matching point to removal, obtain the correct Feature Points Matching point pair of X group；Iteration threshold computing module 210
For being calculated using the following equation iteration threshold r:
Wherein, p is the correct point of expectation to probability, and Sn is the Feature Points Matching point that obtains after nth iteration to sum, Sn=
Sn-1+X-X', S0=0, X' are the Feature Points Matching point that obtains after (n-1)th iteration to the corresponding Feature Points Matching point pair of sum
In, the number of Feature Points Matching point pair identical with the correct Feature Points Matching point centering of X group；Iteration control module 211 is used for
Judge whether iteration threshold r is greater than or equal to 1, if iteration threshold r is greater than or equal to 1, and the number of iterations n is added 1, and call
Registration parameter computing module, authentication module, error hiding removal module and iteration threshold computing module are iterated, wherein n's is first
Initial value is 0, if iteration threshold r less than 1, terminates iteration, the Feature Points Matching point obtained after nth iteration is corresponding to sum
Feature Points Matching point is to for the match point with reference to star chart and star chart subject to registration.
The image registration splicing apparatus based on geometrical invariants provided using the embodiment is chosen and is matched in characteristic point
When introduce geometrical invariants and matched, enable Registration and connection device to have the characteristics that adaptivity, can be good at
Overcome and lack texture information and the difficulty there are ambient noise in star chart, accurately chooses suitable characteristic point and matched,
It is good with accuracy.
Preferably, binarization block 201 specifically executes following steps: will be respectively converted into reference to star chart and star chart subject to registration
Grey level histogram；The corresponding grey level histogram of reference star chart is calculated using least square method, obtains the first optimal threshold,
In, the first optimal threshold keeps the variance with reference to asterism region and background area in star chart maximum；It is corresponding to star chart subject to registration
Grey level histogram is calculated using least square method, obtains the second optimal threshold, wherein the second optimal threshold makes in star chart subject to registration
Asterism region and background area variance it is maximum；Binary conversion treatment is carried out to reference star chart according to the first optimal threshold, is obtained
With reference to M asterism in star chart；Binary conversion treatment is carried out to star chart subject to registration according to the second optimal threshold, obtains star chart subject to registration
In N number of asterism.
Preferably, the mass center of asterism obtains feature on the image after the calculating of characteristic point computing module 202 binarization segmentation
Point.
Preferably, description vectors computing module 204 is calculated using the following equation concentrates in fixed reference feature point, each characteristic point
To the distance metric of remaining characteristic point:
lpi(k)=| | pi-pk||2, k=1,2..., M
Wherein, k ≠ i,For fixed reference feature point set, piAnd pkIt is that fixed reference feature point concentrates characteristic point
Coordinate, lpiIt (k) is piTo pkDistance metric.
It is calculated using the following equation in feature point set subject to registration, the distance metric of each characteristic point to remaining characteristic point:
lqj(l)=| | qj-ql||2, l=1,2 ..., N
Wherein, and l ≠ j,For feature point set subject to registration, qjAnd qlIt is in feature point set subject to registration
The coordinate of characteristic point, lqjIt (l) is qjTo qlDistance metric.
Preferably, in matching value computing module 205, the matching cost function of construction is as follows:
Wherein, δ is the standard deviation of Gaussian function, EijFor piWith qjMatching value.
Preferably, registration parameter computing module 207 becomes each group of Feature Points Matching point to calculating rotation, translation, scale
The parameter with projective transformation is changed, obtains each group of Feature Points Matching point to corresponding registration parameter.
Preferably, authentication module 208 is verified using N2 group Feature Points Matching point to registration parameter, if error is less than
Default error threshold, then it is assumed that this group of Feature Points Matching point is correct.
Embodiment three
The present embodiment also provides a kind of computer equipment, can such as execute the smart phone, tablet computer, notebook of program
Computer, desktop computer, rack-mount server, blade server, tower server or Cabinet-type server are (including independent
Server cluster composed by server or multiple servers) etc..Fig. 3 is that the computer that the embodiment of the present invention four provides is set
Standby hardware structure diagram, as shown in figure 3, include, but is not limited to: can be total by system for the computer equipment 20 of the present embodiment
Line is in communication with each other the memory 21 of connection, processor 22, as shown in Figure 3.It should be pointed out that Fig. 3 is illustrated only with component
The computer equipment 20 of 21-22, it should be understood that being not required for implementing all components shown, the implementation that can be substituted
More or less component.
In the present embodiment, memory 21 (i.e. readable storage medium storing program for executing) includes flash memory, hard disk, multimedia card, card-type memory
(for example, SD or DX memory etc.), random access storage device (RAM), static random-access memory (SRAM), read-only memory
(ROM), electrically erasable programmable read-only memory (EEPROM), programmable read only memory (PROM), magnetic storage, magnetic
Disk, CD etc..In some embodiments, memory 21 can be the internal storage unit of computer equipment 20, such as the calculating
The hard disk or memory of machine equipment 20.In further embodiments, memory 21 is also possible to the external storage of computer equipment 20
The plug-in type hard disk being equipped in equipment, such as the computer equipment 20, intelligent memory card (Smart Media Card, SMC), peace
Digital (Secure Digital, SD) card, flash card (Flash Card) etc..Certainly, memory 21 can also both include meter
The internal storage unit for calculating machine equipment 20 also includes its External memory equipment.In the present embodiment, memory 21 is commonly used in storage
It is installed on the operating system and types of applications software of computer equipment 20, such as the image based on geometrical invariants of embodiment two
The program code etc. of Registration and connection device.In addition, memory 21 can be also used for temporarily storing and export or will be defeated
Various types of data out.
Processor 22 can be in some embodiments central processing unit (Central Processing Unit, CPU),
Controller, microcontroller, microprocessor or other data processing chips.The processor 22 is commonly used in control computer equipment
20 overall operation.In the present embodiment, program code or processing data of the processor 22 for being stored in run memory 21,
Such as image registration splicing apparatus based on geometrical invariants etc..
Example IV
The present embodiment also provides a kind of computer readable storage medium, such as flash memory, hard disk, multimedia card, card-type memory
(for example, SD or DX memory etc.), random access storage device (RAM), static random-access memory (SRAM), read-only memory
(ROM), electrically erasable programmable read-only memory (EEPROM), programmable read only memory (PROM), magnetic storage, magnetic
Disk, CD, server, App are stored thereon with computer program, phase are realized when program is executed by processor using store etc.
Answer function.The computer readable storage medium of the present embodiment is used to store the image registration splicing apparatus based on geometrical invariants,
The image registration joining method based on geometrical invariants of embodiment one is realized when being executed by processor.
Through the foregoing embodiment it is found that the image registration joining method provided by the invention based on geometrical invariants, at least
Realize it is following the utility model has the advantages that
Image registration joining method based on geometrical invariants introduces geometrical invariants when characteristic point is chosen and matches
It is matched, Registration and connection method is enabled to have the characteristics that adaptivity, can be good at overcoming and lack texture in star chart
Information and difficulty there are ambient noise, accurately choose suitable characteristic point and are matched, and matching accuracy is good.
Although some specific embodiments of the invention are described in detail by example, the skill of this field
Art personnel it should be understood that example above merely to being illustrated, the range being not intended to be limiting of the invention.The skill of this field
Art personnel are it should be understood that can without departing from the scope and spirit of the present invention modify to above embodiments.This hair
Bright range is defined by the following claims.
Claims (10)
1. a kind of image registration joining method based on geometrical invariants characterized by comprising
Step S1: carrying out binary conversion treatment to reference star chart and star chart subject to registration respectively, obtains the M star with reference in star chart
N number of asterism in point and the star chart subject to registration, wherein M and N is the natural number more than or equal to 4；
Step S2: the corresponding characteristic point of each asterism is calculated；
Step S3: the coordinate of each characteristic point is calculated, wherein the corresponding spy of the M asterism with reference in star chart
The coordinate of sign point constitutes fixed reference feature point set, the coordinate structure of the corresponding characteristic point of the N number of asterism in the star chart subject to registration
At feature point set subject to registration；
Step S4: concentrating in the fixed reference feature point, calculates each characteristic point to the distance metric of remaining characteristic point, obtains described
Fixed reference feature point concentrates the geometrical invariants description vectors of each characteristic point, in the feature point set subject to registration, calculates each
Characteristic point to the distance metric of remaining characteristic point, retouch by the geometrical invariants for obtaining each characteristic point in the feature point set subject to registration
State vector, wherein it includes M-1 apart from degree that the fixed reference feature point, which concentrates the geometrical invariants description vectors of each characteristic point,
It measures, the geometrical invariants description vectors of each characteristic point include N-1 distance metric in the feature point set subject to registration；
Step S5: by the geometrical invariants description vectors of the fixed reference feature point one characteristic point of concentration and the feature subject to registration
Point concentrates the geometrical invariants description vectors an of characteristic point to input preset matching cost function, obtains of two characteristic points
With value；
Step S6: judging whether the matching value is greater than preset matching threshold value, when the matching value is greater than the preset matching threshold
When value, corresponding two characteristic points of the matching value are one group of Feature Points Matching point pair, obtain S group Feature Points Matching point pair；
Step S7: the N1 group Feature Points Matching point pair, the spy described in each group are extracted in the S group Feature Points Matching point pair
Sign point matching double points are calculated according to rigid body translation model respectively, obtain each group described in Feature Points Matching point match to corresponding
Quasi- parameter；
Step S8: for Feature Points Matching point described in each group to corresponding registration parameter, it is utilized respectively the N2 group characteristic point
Whether matching double points verify the registration parameter, to determine this group of Feature Points Matching point to correct, wherein N1+N2=
S, and the N2 group Feature Points Matching point to be the S group Feature Points Matching point centering except the N1 group Feature Points Matching point to it
Outer Feature Points Matching point pair；
Step S9: by the incorrect Feature Points Matching point of N1 group Feature Points Matching point centering to removal, it is correct to obtain X group
Feature Points Matching point pair；
Step S10: it is calculated using the following equation iteration threshold r:
Wherein, p is the correct point of expectation to probability, SnIt is the Feature Points Matching point that is obtained after nth iteration to sum,
Sn=Sn-1+ X-X', S0=0, X' are the Feature Points Matching point that obtains after (n-1)th iteration to the corresponding Feature Points Matching point of sum
Centering, the number of Feature Points Matching point pair identical with the correct Feature Points Matching point centering of the X group；
Step S11: judging whether the iteration threshold r is greater than or equal to 1, if the iteration threshold r is greater than or equal to 1, simultaneously
The number of iterations n being added 1, return step S7, wherein the initial value of n is 0, if the iteration threshold r terminates iteration less than 1,
The Feature Points Matching point obtained after nth iteration to the corresponding Feature Points Matching point of sum to for it is described refer to star chart and it is described to
It is registrated the match point of star chart.
2. the image registration joining method based on geometrical invariants according to claim 1, which is characterized in that the step S1
Include:
Grey level histogram is respectively converted into reference to star chart and the star chart subject to registration by described；
It is calculated using least square method with reference to the corresponding grey level histogram of star chart described, obtains the first optimal threshold, wherein institute
Stating the first optimal threshold keeps the variance with reference to the asterism region and background area in star chart maximum；
The corresponding grey level histogram of the star chart subject to registration is calculated using least square method, obtains the second optimal threshold, wherein
Second optimal threshold keeps the variance in asterism region and background area in the star chart subject to registration maximum；
Binary conversion treatment is carried out with reference to star chart to described according to first optimal threshold, obtains M with reference in star chart
Asterism；
Binary conversion treatment is carried out to the star chart subject to registration according to second optimal threshold, is obtained in the star chart subject to registration
N number of asterism.
3. the image registration joining method based on geometrical invariants according to claim 1, which is characterized in that the step S2
Include:
The mass center of the asterism obtains the characteristic point on image after calculating binarization segmentation.
4. the image registration joining method based on geometrical invariants according to claim 1, which is characterized in that in the step
In S4:
It is calculated using the following equation and is concentrated in the fixed reference feature point, the distance metric of each characteristic point to remaining characteristic point:
lpi(k)=| | pi-pk||2, k=1,2..., M
Wherein, k ≠ i,For the fixed reference feature point set, piAnd pkIt is that the fixed reference feature point concentrates spy
Levy the coordinate of point, lpiIt (k) is piTo pkDistance metric；
It is calculated using the following equation in the feature point set subject to registration, the distance metric of each characteristic point to remaining characteristic point:
lqj(l)=| | qj-ql||2, l=1,2..., N
Wherein, and l ≠ j,For the feature point set subject to registration, qjAnd qlIt is the characteristic point subject to registration
Concentrate the coordinate of characteristic point, lqjIt (l) is qjTo qlDistance metric.
5. the image registration joining method based on geometrical invariants according to claim 4, which is characterized in that the matching of construction
Cost function is as follows:
Wherein, δ is the standard deviation of Gaussian function, EijFor piWith qjMatching value.
6. the image registration joining method based on geometrical invariants according to claim 1, which is characterized in that the step S7
Include:
The point of the Feature Points Matching described in each group obtains each to calculating rotation, translating, the parameter of change of scale and projective transformation
The group Feature Points Matching point is to corresponding registration parameter.
7. the image registration joining method based on geometrical invariants according to claim 1, which is characterized in that the step S8
Include:
It is verified using the N2 group Feature Points Matching point to the registration parameter, if error is less than default error threshold,
Then think that this group of Feature Points Matching point is correct.
8. a kind of image registration splicing apparatus based on geometrical invariants characterized by comprising
Binarization block obtains described with reference to star chart for carrying out binary conversion treatment respectively to reference star chart and star chart subject to registration
In M asterism and the star chart subject to registration in N number of asterism, wherein M and N is the natural number more than or equal to 4；
Characteristic point computing module, for calculating the corresponding characteristic point of each asterism；
Coordinate calculation module, for calculating the coordinate of each characteristic point, wherein the M star with reference in star chart
The coordinate of the corresponding characteristic point of point constitutes fixed reference feature point set, the corresponding feature of the N number of asterism in the star chart subject to registration
The coordinate of point constitutes feature point set subject to registration；
Description vectors computing module, for being concentrated in the fixed reference feature point, calculate each characteristic point to remaining characteristic point away from
From measurement, the geometrical invariants description vectors that the fixed reference feature point concentrates each characteristic point are obtained, in the feature subject to registration
Point is concentrated, and is calculated each characteristic point to the distance metric of remaining characteristic point, is obtained each feature in the feature point set subject to registration
The geometrical invariants description vectors of point, wherein the fixed reference feature point concentrates the geometrical invariants description vectors of each characteristic point
Including M-1 distance metric, the geometrical invariants description vectors of each characteristic point include N-1 in the feature point set subject to registration
Distance metric；
Matching value computing module, for the fixed reference feature point to be concentrated to the geometrical invariants description vectors and institute of a characteristic point
The geometrical invariants description vectors for stating a characteristic point in feature point set subject to registration input preset matching cost function, obtain two
The matching value of a characteristic point；
Matching double points determining module, for judging whether the matching value is greater than preset matching threshold value, when the matching value is greater than
When the preset matching threshold value, corresponding two characteristic points of the matching value are one group of Feature Points Matching point pair, obtain S group feature
Point matching double points；
Registration parameter computing module, for extracting the N1 group Feature Points Matching point pair in the S group Feature Points Matching point pair,
The point of the Feature Points Matching described in each group to calculating respectively according to rigid body translation model, obtain each group described in characteristic point
With point to corresponding registration parameter；
Authentication module, for, to corresponding registration parameter, being utilized respectively described in N2 group for Feature Points Matching point described in each group
Whether Feature Points Matching point is verified to the registration parameter, to determine this group of Feature Points Matching point to correct, wherein N1
+ N2=S, and the N2 group Feature Points Matching point removes the N1 group Feature Points Matching to for the S group Feature Points Matching point centering
Feature Points Matching point pair of the point to except；
Error hiding removes module, for by the incorrect Feature Points Matching point of N1 group Feature Points Matching point centering to removal,
Obtain the correct Feature Points Matching point pair of X group；
Iteration threshold computing module, for being calculated using the following equation iteration threshold r:
Wherein, p is the correct point of expectation to probability, SnIt is the Feature Points Matching point that is obtained after nth iteration to sum,
Sn=Sn-1+ X-X', S0=0, X' are the Feature Points Matching point that obtains after (n-1)th iteration to the corresponding Feature Points Matching point of sum
Centering, the number of Feature Points Matching point pair identical with the correct Feature Points Matching point centering of the X group；
Iteration control module, for judging whether the iteration threshold r is greater than or equal to 1, if the iteration threshold r is greater than or waits
In 1, then and by the number of iterations n add 1, and calls the registration parameter computing module, the authentication module, error hiding removal
Module and iteration threshold computing module are iterated, wherein the initial value of n is 0, if the iteration threshold r terminates less than 1
Iteration, the Feature Points Matching point obtained after nth iteration to the corresponding Feature Points Matching point of sum to for it is described with reference to star chart with
The match point of the star chart subject to registration.
9. a kind of computer equipment, the computer equipment include memory, processor and storage on a memory and can be
The computer program run on processor, which is characterized in that the processor realizes claim 1 to 7 when executing described program
The step of any one the method.
10. a kind of computer readable storage medium, is stored thereon with computer program, it is characterised in that: described program is processed
The step of any one of claim 1 to 7 the method is realized when device executes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810730583.1A CN108961164A (en) | 2018-07-05 | 2018-07-05 | Image registration joining method and device and computer equipment based on geometrical invariants |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810730583.1A CN108961164A (en) | 2018-07-05 | 2018-07-05 | Image registration joining method and device and computer equipment based on geometrical invariants |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108961164A true CN108961164A (en) | 2018-12-07 |
Family
ID=64484212
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810730583.1A Pending CN108961164A (en) | 2018-07-05 | 2018-07-05 | Image registration joining method and device and computer equipment based on geometrical invariants |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108961164A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110097533A (en) * | 2019-02-12 | 2019-08-06 | 哈尔滨新光光电科技股份有限公司 | A kind of method for accurate testing of hot spot outer dimension and position |
| CN111091590A (en) * | 2019-12-18 | 2020-05-01 | Oppo广东移动通信有限公司 | Image processing method, image processing device, storage medium and electronic equipment |
| CN113298713A (en) * | 2021-05-24 | 2021-08-24 | 北京市遥感信息研究所 | On-orbit rapid registration method capable of resisting cloud interference |
| CN114820738A (en) * | 2022-06-30 | 2022-07-29 | 中国人民解放军国防科技大学 | Accurate registration method and device for star atlas, computer equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104732546A (en) * | 2015-04-02 | 2015-06-24 | 西安电子科技大学 | Non-rigid SAR image registration method based on region similarity and local spatial constraint |
| CN105631870A (en) * | 2015-12-25 | 2016-06-01 | 北京理工大学 | Star image registration and target track extraction method based on NoC framework and device thereof |
| CN106846386A (en) * | 2017-02-08 | 2017-06-13 | 南通大学 | 3D cranium method for registering images based on ROI and conformal geometric algebra property invariant |
-
2018
- 2018-07-05 CN CN201810730583.1A patent/CN108961164A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104732546A (en) * | 2015-04-02 | 2015-06-24 | 西安电子科技大学 | Non-rigid SAR image registration method based on region similarity and local spatial constraint |
| CN105631870A (en) * | 2015-12-25 | 2016-06-01 | 北京理工大学 | Star image registration and target track extraction method based on NoC framework and device thereof |
| CN106846386A (en) * | 2017-02-08 | 2017-06-13 | 南通大学 | 3D cranium method for registering images based on ROI and conformal geometric algebra property invariant |
Non-Patent Citations (1)
| Title |
|---|
| 庄幽文 等: "基于几何不变量的形状上下文星图配准", 《光学精密工程》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110097533A (en) * | 2019-02-12 | 2019-08-06 | 哈尔滨新光光电科技股份有限公司 | A kind of method for accurate testing of hot spot outer dimension and position |
| CN110097533B (en) * | 2019-02-12 | 2023-04-07 | 哈尔滨新光光电科技股份有限公司 | Method for accurately testing overall dimension and position of light spot |
| CN111091590A (en) * | 2019-12-18 | 2020-05-01 | Oppo广东移动通信有限公司 | Image processing method, image processing device, storage medium and electronic equipment |
| CN111091590B (en) * | 2019-12-18 | 2024-04-23 | Oppo广东移动通信有限公司 | Image processing method, device, storage medium and electronic equipment |
| CN113298713A (en) * | 2021-05-24 | 2021-08-24 | 北京市遥感信息研究所 | On-orbit rapid registration method capable of resisting cloud interference |
| CN114820738A (en) * | 2022-06-30 | 2022-07-29 | 中国人民解放军国防科技大学 | Accurate registration method and device for star atlas, computer equipment and storage medium |
| CN114820738B (en) * | 2022-06-30 | 2022-09-23 | 中国人民解放军国防科技大学 | Accurate registration method and device for star atlas, computer equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Aldoma et al. | Multimodal cue integration through hypotheses verification for rgb-d object recognition and 6dof pose estimation | |
| CN108961164A (en) | Image registration joining method and device and computer equipment based on geometrical invariants | |
| CN109376631B (en) | Loop detection method and device based on neural network | |
| CN110287873B (en) | Non-cooperative target pose measurement method and system based on deep neural network and terminal equipment | |
| Cicconet et al. | A convolutional approach to reflection symmetry | |
| CN111459269B (en) | Augmented reality display method, system and computer readable storage medium | |
| US20220392201A1 (en) | Image feature matching method and related apparatus, device and storage medium | |
| Heinly et al. | Correcting for duplicate scene structure in sparse 3D reconstruction | |
| CN109348731A (en) | A kind of method and device of images match | |
| CN105934757A (en) | Method and apparatus for detecting incorrect associations between keypoints of first image and keypoints of second image | |
| Liang et al. | Image registration using two-layer cascade reciprocal pipeline and context-aware dissimilarity measure | |
| Jog et al. | Automated computation of the fundamental matrix for vision based construction site applications | |
| Shi et al. | Robust image registration using structure features | |
| CN109840529B (en) | Image matching method based on local sensitivity confidence evaluation | |
| Zhou et al. | Vision-based pose estimation from points with unknown correspondences | |
| Li et al. | Guided neighborhood affine subspace embedding for feature matching | |
| Zakharov et al. | Finding correspondences between images using descriptors and graphs | |
| CN105190689A (en) | Image processing including adjoin feature based object detection, and/or bilateral symmetric object segmentation | |
| Wang et al. | Common visual pattern discovery via nonlinear mean shift clustering | |
| Lim et al. | Estimation of the epipole using optical flow at antipodal points | |
| Wen et al. | An improved SIFT operator-based image registration using cross-correlation information | |
| Wang et al. | Unified detection of skewed rotation, reflection and translation symmetries from affine invariant contour features | |
| CN116958021A (en) | Product defect identification method based on artificial intelligence, related device and medium | |
| CN115393903A (en) | Method, device and equipment for updating image base and storage medium | |
| Je et al. | Homographic p-norms: Metrics of homographic image transformation |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181207 |