CN110427966A - One kind rejecting error hiding feature point methods based on characteristic point local feature - Google Patents
One kind rejecting error hiding feature point methods based on characteristic point local feature Download PDFInfo
- Publication number
- CN110427966A CN110427966A CN201910558251.4A CN201910558251A CN110427966A CN 110427966 A CN110427966 A CN 110427966A CN 201910558251 A CN201910558251 A CN 201910558251A CN 110427966 A CN110427966 A CN 110427966A
- Authority
- CN
- China
- Prior art keywords
- point
- characteristic point
- feature
- matching
- thick
- 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 41
- 238000012360 testing method Methods 0.000 claims abstract description 16
- 238000001514 detection method Methods 0.000 claims abstract description 14
- 238000012887 quadratic function Methods 0.000 claims description 4
- 239000000284 extract Substances 0.000 claims 1
- 239000012141 concentrate Substances 0.000 abstract description 4
- 239000011159 matrix material Substances 0.000 description 5
- 238000010276 construction Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 description 1
- HUTDUHSNJYTCAR-UHFFFAOYSA-N ancymidol Chemical compound C1=CC(OC)=CC=C1C(O)(C=1C=NC=NC=1)C1CC1 HUTDUHSNJYTCAR-UHFFFAOYSA-N 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004907 flux Effects 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 238000010998 test method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses one kind to reject error hiding feature point methods based on characteristic point local feature comprising following steps: SIFT feature detection carries out extreme point detection using the Gaussian scale-space of building, obtains key point；Feature Points Matching, the key point descriptor information based on acquisition calculate minimum distance and time short distance in point distance the second width image in piece image using Euclidean distance, and minimum distance is smaller, then matching degree is higher；Error hiding feature point methods are rejected based on characteristic point local feature value, uses the ratio testing of lenient thresholds to reject to be thick, rejects a small amount of error hiding characteristic point, guarantee enough correct characteristic points of matching；Local coordinate system is constructed in two width matching images respectively, concentrates matched characteristic point in the similarity of local coordinate system lower eigenvalue according to, rejects error hiding, realize that essence is rejected.The present invention effectively rejects the error hiding in SIFT algorithmic match result, can get the high point set of accuracy.
Description
Technical field
The present invention relates to a kind of methods for rejecting error hiding characteristic point, more particularly, to based on characteristic point local feature value
The method for rejecting error hiding characteristic point.
Background technique
Feature Points Matching is the basis much applied in computer vision field, in image registration, image mosaic, Three-dimensional Gravity
It founds a capital, play an important role in target identification.In three-dimensional reconstruction, the correct of matching double points directly decides camera projection matrix
Solution, and the calculating of three-dimensional point coordinate is carried out according to matching characteristic point and camera projection matrix, so characteristic point
Matching directly determine the precision of threedimensional model, during three-dimensional reconstruction, the matching of characteristic point plays decisive role.In
When image mosaic, image registration, need to solve the registration of the geometric transformation model realization image of image according to matched characteristic point,
The accuracy of Feature Points Matching decides image mosaic, the precision of registration.
Three steps can be divided into using Feature Points Matching: characteristic point detection, construction feature point describe son, feature point description carries out
Matching.Common feature point extraction algorithm has SIFT, SUFR, ORB, Harris etc., and SIFT algorithm is due to image scaling, rotation
Turn and affine transformation all has good invariance, stability is strong, matching precision is high and is widely applied；But it is matching
When, the application of algorithm is influenced there are will cause error hiding when similar grain region since image lacks texture information.For
The characteristic point of error hiding, commonly used method are RANSAC algorithm (RANSAC) and its a series of improved calculations
Method selects the matching for meeting particular matrix according to basis matrix or homography matrix by being iterated to initial point set
The most point set of point is as final matching point set, therefore the speed of the more high then iteration of the initial matched accuracy of point set is faster,
Meanwhile when the characteristic point of initial point concentration error hiding is more, apparent error hiding characteristic point is still had using RANSAC and is not had
Have and is effectively rejected.Therefore initial point set is handled, it is meaningful for improving matching accuracy.
Initial matching point set, the i.e. ratio of the arest neighbors of feature vector and time neighbour are obtained using ratio testing in conventional method
It is then same place when value is less than threshold value, but matched result is affected by threshold value, when threshold value is excessively loose, it may appear that mistake
It omits with, does not reject effectively, threshold value is excessively tight, can will match correct characteristic point and reject.
Therefore, the prior art needs further improvement and develops.
Summary of the invention
In view of above-mentioned deficiencies of the prior art, the present invention for this problem, is proposed and is picked based on characteristic point local feature value
Except error hiding feature method, this method includes that thick reject rejects two processes with essence.Lenient thresholds are used during thick reject
Ratio testing be it is thick reject, reject a small amount of error hiding characteristic point, while guaranteeing enough correct characteristic points of matching；It is rejected in essence
In the process, similar local coordinate system is constructed in two width matching images respectively, concentrates matched characteristic point in local seat according to
The similarity of mark system lower eigenvalue rejects error hiding, realizes that essence is rejected.
In order to solve the above technical problems, the present invention program includes:
One kind rejecting error hiding feature point methods based on characteristic point local feature comprising following steps:
A, SIFT feature detects, and does the Fuzzy Processing of different scale parameter using Gaussian convolution collecting image and drop is adopted
Sample constructs Gaussian scale-space, and figure layer adjacent in Gaussian scale-space is subtracted each other building difference Gaussian scale-space；In Gaussian difference
By pixel, totally 26 points are compared with the point of 8 neighborhoods, two neighboring level 9 × 2 in point space, carry out extreme point detection,
Obtain key point.
B, Feature Points Matching, the key point descriptor information based on acquisition are calculated in piece image using Euclidean distance
Point distance the second width image in minimum distance and time closely, minimum distance is smaller, then matching degree is higher, therefore with nearest
It is thick matching characteristic point set apart from matched result, is then rejected using ratio testing method.
C, error hiding feature point methods are rejected based on characteristic point local feature value, this method includes that thick rejecting and essence reject two
A process, it is thick to reject, it uses the ratio testing of lenient thresholds to reject to be thick, rejects a small amount of error hiding characteristic point, while guaranteeing foot
Flux matched correct characteristic point；Essence is rejected, and constructs similar local coordinate system in two width matching images respectively, according to a concentration
The characteristic point matched rejects error hiding in the similarity of local coordinate system lower eigenvalue, realizes that essence is rejected.
The feature matching method, wherein step A is specific further include: in order to obtain stable key point, using three
Dimension quadratic function is fitted key point, removes the unstable point of low contrast and edge.
The feature matching method, wherein the step C specifically includes: thick rejecting, for one group of image I1(x,
And I y)2(x, y) carries out characteristic point detection using SIFT algorithm, and the group that uses force lifts algorithm and calculates thick matching characteristic point set,
Based on the point set, fixed threshold is used to handle for the ratio testing method of Thred1 thick matching point set, guarantee has enough
When matching correct characteristic point, part error hiding characteristic point is rejected.
Local rectangular coordinate system is constructed, it is correct to filter out 3 pairs of matchings from the point concentration after thick reject in order to quickly and accurately
Characteristic point construct orthogonal coordinate system, in image I1(x, y) and I2It is constructed in (x, y) by distance restraint and angle restriction
Similar triangles are extracted by triangle similarity principle and match correct characteristic point.
Characteristic point local feature value is calculated, thick coordinate of the matched characteristic point in its local coordinate system, calculating are calculated
Euclidean distance with point between, it is contemplated that the calculating error of coordinate, then when Euclidean distance is less than specific threshold Thred2, i.e.,
To match correct characteristic point.
It is provided by the invention a kind of based on characteristic point local feature value rejecting error hiding feature point methods.This method is with high threshold
The result that value ratio testing obtains is initial matching point set, is based on triangle similarity principle, filters out 3 from this feature point concentration
A correct characteristic point pair of matching, constructs local rectangular coordinate system in benchmark image and measuring image respectively using it, according to
Matched characteristic point rejects error hiding to the similarity of the local feature value under different coordinates；The present invention effectively rejects SIFT
Error hiding in algorithmic match result, accuracy is higher, reduces and matches the probability that correct characteristic point is accidentally rejected, and can get
The high point set of accuracy.
Detailed description of the invention
Fig. 1 is the flow diagram that characteristic point local feature rejects error hiding feature point methods in the present invention.
Specific embodiment
The present invention provides a kind of characteristic point local feature reject error hiding feature point methods, for make the purpose of the present invention,
Technical solution and effect are clearer, clear, and the present invention is described in more detail below.It should be appreciated that described herein
Specific embodiment is only used to explain the present invention, is not intended to limit the present invention.
The present invention provides a kind of characteristic point local features to reject error hiding feature point methods, as shown in Figure 1 comprising
Following steps:
Step 101:SIFT characteristic point detection, using Gaussian convolution collecting image do the Fuzzy Processing of different scale parameter with
And down-sampled building Gaussian scale-space, figure layer adjacent in Gaussian scale-space is subtracted each other into building difference Gaussian scale-space；In
By pixel, totally 26 points are compared with the point of 8 neighborhoods, two neighboring level 9 × 2 in difference of Gaussian space, carry out extreme point
Detection is obtained key point and is fitted using three-dimensional quadratic function to key point to obtain stable key point, removed low
The unstable point of contrast and edge；
Step 102: Feature Points Matching, the key point descriptor information based on acquisition calculate the first width using Euclidean distance
Minimum distance and time short distance in point distance the second width image in image, minimum distance is smaller, then matching degree is higher, therefore
With the matched result of minimum distance for thick matching characteristic point set, then rejected using ratio testing method；It calculates nearest
The ratio of distance and secondary short distance:
When T is less than fixed threshold, then it is assumed that this is matching characteristic point to point, and wherein ND is minimum distance, and NND is secondary close
Distance.
Step 103: the process includes that thick reject rejects two processes with essence, larger using threshold value during thick reject
Ratio testing method, retain the correct characteristic point of enough matchings, based on the point set after thick reject, select 3 matchings correct
Characteristic point pair establishes local similar coordinate systems by Smith's orthogonal transformation in two images respectively, and calculating is thick to reject point set
In coordinate of the matching characteristic point in local coordinate system existed using coordinate construction feature value according to correct characteristic point is matched
The principle that characteristic value is equal in coordinate system rejects error hiding characteristic point, achievees the purpose that essence rejecting.
In another preferred embodiment of the present invention, the step 101 specifically includes: the key point that above procedure obtains
With scale invariability, but do not have rotational invariance, therefore is distributed using characteristic point of the local feature of image to extraction
One reference direction, as the principal direction of key point, 80% direction that peak value is greater than principal direction is maximum value using in histogram
Auxiliary direction, at the same in key point scale space in 4 × 4 window calculate 8 directions gradient information, building 128 dimension spies
Vector is levied, and building descriptor information is normalized to feature vector.
Further, the step 103 is specific further include: building local rectangular coordinate system, in order to quickly and accurately
3 pairs of matchings correct characteristic point building orthogonal coordinate system is filtered out from the point concentration after thick reject, in image I1(x, y) and
I2Similar triangles are constructed by distance restraint and angle restriction in (x, y), matching is being extracted just by triangle similarity principle
True characteristic point.Postulated point concentrates matched characteristic point to for { pi, pj, pkIt is I1Characteristic point in (x, y), { pi', pj',
pk' it is I1Characteristic point in (x, y) is in I2Match point in (x, y).
On the basis of above-mentioned, in another preferred embodiment of the invention, the step 103 includes: specifically in coordinate
It is to calculate separately thick match point in O-XY and O '-X ' Y ' and concentrate matched characteristic point to { (pr, pr'), r=1,2,3....N } In
Coordinate in respective coordinates system, in I1In (x, y), coordinate of its characteristic point in O-XY is calculated.
Further, the step 103 specifically includes:
X in formulaA、YA、XB、YBIt is the coordinate information of two substrates of coordinate system O-XY, α, β are characteristic point pr(Xr, Yr) sitting
P can be accurately calculated by (1) formula in coordinate in mark system O-XYrCoordinate α, β in O-XY coordinate system, calculate simultaneously
pr' coordinate α ', the β ' in O '-X ' Y ' coordinate system, if prAnd pr' it is the correct characteristic point pair of matching, then Europe between them
Formula distance:
It is zero, it is contemplated that the calculating error of coordinate is then to match correctly when Euclidean distance is less than specific threshold Tred2
Characteristic point.
More specifically but, the step 103 specifically include: but using Euclidean distance calculate two characteristic points it
Between similarity calculation amount it is big, in order to improve its computational efficiency, introduced feature value F, F are the integers using characteristic point coordinate α, β
Part constructs a real-coded GA, form are as follows:
F=μ σ (3)
Wherein, μ is the integer part of α, and σ is the integer part of β, replaces characteristic point coordinate to calculate similarity using Index
Its computational efficiency can be effectively improved.
In order to which the present invention program is further described, it is exemplified below more detailed embodiment and is illustrated.
The first step, SIFT feature detection, this method detection process mainly include the detection of scale space extreme point, key point
It is accurately positioned, the matching of key point direction and crucial point feature description, SIFT algorithm make different rulers using Gaussian convolution collecting image
It is poor to be subtracted each other building by the Fuzzy Processing of degree parameter and down-sampled building Gaussian scale-space for figure layer adjacent in Gaussian scale-space
Divide Gaussian scale-space；The point of pixel and 8 neighborhoods, two neighboring level 9 × 2 are clicked through for 26 totally in difference of Gaussian space
Row compares, and carries out extreme point detection, key point is obtained, in order to obtain stable key point, using three-dimensional quadratic function to key
Point is fitted, and removes the unstable point of low contrast and edge.The key point that above procedure obtains has scale invariability,
But do not have rotational invariance, therefore distribute a reference direction using characteristic point of the local feature of image to extraction, with
Principal direction of the maximum value as key point in histogram, peak value are greater than direction supplemented by 80% direction of principal direction, while closing
The gradient information for calculating 8 directions in key point scale space in 4 × 4 window constructs the feature vector of 128 dimensions, and to feature
Vector normalization building descriptor information.
Second step, Feature Points Matching, the key point descriptor information based on acquisition calculate the first width figure using Euclidean distance
Minimum distance and time short distance in point distance the second width image as in, minimum distance is smaller, then matching degree is higher, therefore with
The matched result of minimum distance is thick matching characteristic point set, is then rejected using ratio testing method, that is, calculates most low coverage
From the ratio with secondary short distance:
When T is less than fixed threshold, then it is assumed that this is matching characteristic point to point, and wherein ND is minimum distance, and NND is secondary close
Distance.
However, matched result is related with fixed threshold using ratio testing, and when threshold value is larger, matched characteristic point
Quantity is more, but there are a large amount of Mismatching points, when threshold value is smaller, can will match correct characteristic point and accidentally reject, meanwhile, Ke Nengyi
The point that can so have a large amount of error hidings is not rejected effectively.Therefore, the invention proposes picked based on characteristic point local feature value
Except error hiding method, when rejecting error hiding characteristic point, reduces and match correct characteristic point by accidentally rejecting phenomenon.
Third step rejects error hiding feature point methods based on characteristic point local feature value, and this method includes thick rejecting and essence
Reject two processes.During thick reject, using the biggish ratio testing method of threshold value, retain enough correct features of matching
Point is selected 3 correct characteristic points pair of matching, is being passed through Smith just in two images respectively based on the point set after thick reject
Alternation changes, and establishes local similar coordinate systems, and coordinate of the matching characteristic point that the thick rejecting point of calculating is concentrated in local coordinate system is sharp
With coordinate construction feature value, according to correct characteristic point is matched, the equal principle of characteristic value rejects error hiding feature in a coordinate system
Point achievees the purpose that essence rejecting.It is thick to reject, for one group of image I1(x, y) and I2(x, y) carries out characteristic point using SIFT algorithm
Detection, and the group that uses force lifts algorithm and calculates thick matching characteristic point set, the point set is based on, using the ratio of fixed threshold Thred1
Test method handles thick matching point set, when guarantee has enough matchings correct characteristic point, rejects part error hiding feature
Point.Local rectangular coordinate system is constructed, it is correctly special to filter out 3 pairs of matchings from the point concentration after thick reject in order to quickly and accurately
Sign point building orthogonal coordinate system, in image I1(x, y) and I2It is similar with angle restriction building by distance restraint in (x, y)
Triangle is extracted by triangle similarity principle and matches correct characteristic point.Characteristic point local feature value is calculated, calculates thick
Coordinate of the characteristic point matched in its local coordinate system calculates the Euclidean distance between matching double points, it is contemplated that the calculating of coordinate
Error is then to match correct characteristic point when Euclidean distance is less than specific threshold.
4th step checks matching effect.The present invention can effectively reject the error hiding in SIFT algorithmic match result, together
When, compared with testing matching result with low-ratio, accuracy is higher, reduces and matches the probability that correct characteristic point is accidentally rejected.
Because the present invention can effectively reject error hiding characteristic point, the high initial matching point set of accuracy is obtained.
Certainly, described above is only that presently preferred embodiments of the present invention is answered the present invention is not limited to enumerate above-described embodiment
When explanation, anyone skilled in the art is all equivalent substitutes for being made, bright under the introduction of this specification
Aobvious variant, all falls within the essential scope of this specification, ought to be by protection of the invention.
Claims (5)
1. a kind of method for rejecting error hiding characteristic point based on characteristic point local feature comprising following steps:
A, SIFT feature detect, using Gaussian convolution collecting image do different scale parameter Fuzzy Processing and down-sampled structure
Gaussian scale-space is built, figure layer adjacent in Gaussian scale-space is subtracted each other into building difference Gaussian scale-space；In difference of Gaussian sky
Between in by pixel, totally 26 points are compared with the point of 8 neighborhoods, two neighboring level 9 × 2, carry out extreme point detection, obtain
Key point.
B, Feature Points Matching, the key point descriptor information based on acquisition calculate the point in piece image using Euclidean distance
Minimum distance and time short distance in distance the second width image, minimum distance is smaller, then matching degree is higher, therefore with minimum distance
Matched result is thick matching characteristic point set, is then rejected using ratio testing method.
C, error hiding feature point methods are rejected based on characteristic point local feature value, this method includes that thick rejecting and essence reject two mistakes
Journey, it is thick to reject, it uses the ratio testing of lenient thresholds to reject to be thick, rejects a small amount of error hiding characteristic point, while guaranteeing enough
With correct characteristic point；Essence is rejected, and constructs similar local coordinate system in two width matching images respectively, matched according to concentration
Characteristic point rejects error hiding in the similarity of local coordinate system lower eigenvalue, realizes that essence is rejected.
2. feature matching method according to claim 1, which is characterized in that the step A specifically includes: to obtain
Stable key point is fitted key point using three-dimensional quadratic function, removes the unstable point of low contrast and edge.
3. feature matching method according to claim 1, which is characterized in that the step C specifically includes: thick rejecting,
For one group of image I1(x, y) and I2(x, y) carries out characteristic point detection using SIFT algorithm, and the group that uses force lifts algorithm and calculates
Thick matching characteristic point set, is based on the point set, and fixed threshold is used to carry out for the ratio testing method of Thred1 to thick matching point set
Processing when guarantee has enough matchings correct characteristic point, rejects part error hiding characteristic point.
4. feature matching method according to claim 3, which is characterized in that the step C is specific further include: building office
Portion's rectangular coordinate system, in order to quickly and accurately filter out the correct characteristic point building of 3 pairs of matchings just from the point concentration after thick reject
Rectangular coordinate system is handed over, in image I1(x, y) and I2Similar triangles are constructed by distance restraint and angle restriction in (x, y), are passed through
Triangle similarity principle, which extracts, matches correct characteristic point.
5. feature matching method according to claim 3, which is characterized in that the step C is specific further include: calculate special
Sign point local feature value, calculates thick coordinate of the matched characteristic point in its local coordinate system, calculates the Europe between matching double points
Formula distance, it is contemplated that the calculating error of coordinate is then to match correctly spy when Euclidean distance is less than specific threshold Thred2
Sign point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910558251.4A CN110427966A (en) | 2019-06-17 | 2019-06-17 | One kind rejecting error hiding feature point methods based on characteristic point local feature |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910558251.4A CN110427966A (en) | 2019-06-17 | 2019-06-17 | One kind rejecting error hiding feature point methods based on characteristic point local feature |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110427966A true CN110427966A (en) | 2019-11-08 |
Family
ID=68408752
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910558251.4A Pending CN110427966A (en) | 2019-06-17 | 2019-06-17 | One kind rejecting error hiding feature point methods based on characteristic point local feature |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110427966A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111192254A (en) * | 2019-12-30 | 2020-05-22 | 无锡信捷电气股份有限公司 | Weld joint feature point filtering method based on global threshold and template matching |
| CN111401252A (en) * | 2020-03-17 | 2020-07-10 | 广东技术师范大学 | Spine matching method and equipment of book checking system based on vision |
| CN111476251A (en) * | 2020-03-26 | 2020-07-31 | 中国人民解放军战略支援部队信息工程大学 | Remote sensing image matching method and device |
| CN112001954A (en) * | 2020-08-20 | 2020-11-27 | 大连海事大学 | Polar curve constraint-based underwater PCA-SIFT image matching method |
| CN112419314A (en) * | 2020-12-10 | 2021-02-26 | 易思维（杭州）科技有限公司 | Characteristic point eliminating method based on correlation |
| CN113033257A (en) * | 2019-12-24 | 2021-06-25 | 杭州海康威视数字技术股份有限公司 | Fingerprint identification method and device, electronic equipment and readable storage medium |
| CN113160284A (en) * | 2021-03-09 | 2021-07-23 | 大连海事大学 | Guiding space consistent photovoltaic image registration method based on local similar structure constraint |
| CN113221914A (en) * | 2021-04-14 | 2021-08-06 | 河海大学 | Image feature point matching and mismatching elimination method based on Jacobsad distance |
| CN114255051A (en) * | 2021-12-21 | 2022-03-29 | 中科星通(廊坊)信息技术有限公司 | Authenticity inspection method of orthometric product based on stereo mapping satellite |
| CN114777668A (en) * | 2022-04-12 | 2022-07-22 | 新拓三维技术(深圳)有限公司 | Desktop type elbow measuring method and device |
| CN114998773A (en) * | 2022-08-08 | 2022-09-02 | 四川腾盾科技有限公司 | Characteristic mismatching elimination method and system suitable for aerial image of unmanned aerial vehicle system |
| CN117853462A (en) * | 2024-01-11 | 2024-04-09 | 连云港市第二人民医院(连云港市临床肿瘤研究所) | Intra-articular pressure detection and information extraction method and system based on multi-mode imaging |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101819680A (en) * | 2010-05-12 | 2010-09-01 | 上海交通大学 | Detection method of picture matching point pair |
| 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 |
-
2019
- 2019-06-17 CN CN201910558251.4A patent/CN110427966A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101819680A (en) * | 2010-05-12 | 2010-09-01 | 上海交通大学 | Detection method of picture matching point pair |
| 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 |
Non-Patent Citations (2)
| Title |
|---|
| 刘婷婷: "基于单应性矩阵剔除SIFT错误匹配点的方法", 《哈尔滨商业大学学报(自然科学版)》 * |
| 戚海想: "基于仿射变换的局部特征匹配算法", 《现代计算机(专业版)》 * |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113033257B (en) * | 2019-12-24 | 2023-08-18 | 杭州海康威视数字技术股份有限公司 | Fingerprint identification method, fingerprint identification device, electronic equipment and readable storage medium |
| CN113033257A (en) * | 2019-12-24 | 2021-06-25 | 杭州海康威视数字技术股份有限公司 | Fingerprint identification method and device, electronic equipment and readable storage medium |
| CN111192254A (en) * | 2019-12-30 | 2020-05-22 | 无锡信捷电气股份有限公司 | Weld joint feature point filtering method based on global threshold and template matching |
| CN111192254B (en) * | 2019-12-30 | 2022-03-29 | 无锡信捷电气股份有限公司 | Weld joint feature point filtering method based on global threshold and template matching |
| CN111401252A (en) * | 2020-03-17 | 2020-07-10 | 广东技术师范大学 | Spine matching method and equipment of book checking system based on vision |
| CN111476251A (en) * | 2020-03-26 | 2020-07-31 | 中国人民解放军战略支援部队信息工程大学 | Remote sensing image matching method and device |
| CN112001954A (en) * | 2020-08-20 | 2020-11-27 | 大连海事大学 | Polar curve constraint-based underwater PCA-SIFT image matching method |
| CN112001954B (en) * | 2020-08-20 | 2023-09-15 | 大连海事大学 | Underwater PCA-SIFT image matching method based on polar curve constraint |
| CN112419314A (en) * | 2020-12-10 | 2021-02-26 | 易思维（杭州）科技有限公司 | Characteristic point eliminating method based on correlation |
| CN112419314B (en) * | 2020-12-10 | 2023-02-28 | 易思维（杭州）科技有限公司 | Characteristic point eliminating method based on correlation |
| CN113160284B (en) * | 2021-03-09 | 2024-04-30 | 大连海事大学 | Guidance space-consistent photovoltaic image registration method based on local similar structure constraint |
| CN113160284A (en) * | 2021-03-09 | 2021-07-23 | 大连海事大学 | Guiding space consistent photovoltaic image registration method based on local similar structure constraint |
| CN113221914A (en) * | 2021-04-14 | 2021-08-06 | 河海大学 | Image feature point matching and mismatching elimination method based on Jacobsad distance |
| CN113221914B (en) * | 2021-04-14 | 2022-10-11 | 河海大学 | Image feature point matching and mismatching elimination method based on Jacobsad distance |
| CN114255051A (en) * | 2021-12-21 | 2022-03-29 | 中科星通(廊坊)信息技术有限公司 | Authenticity inspection method of orthometric product based on stereo mapping satellite |
| CN114777668B (en) * | 2022-04-12 | 2024-01-16 | 新拓三维技术(深圳)有限公司 | Desktop type bent pipe measuring method and device |
| CN114777668A (en) * | 2022-04-12 | 2022-07-22 | 新拓三维技术(深圳)有限公司 | Desktop type elbow measuring method and device |
| CN114998773B (en) * | 2022-08-08 | 2023-02-17 | 四川腾盾科技有限公司 | Characteristic mismatching elimination method and system suitable for aerial image of unmanned aerial vehicle system |
| CN114998773A (en) * | 2022-08-08 | 2022-09-02 | 四川腾盾科技有限公司 | Characteristic mismatching elimination method and system suitable for aerial image of unmanned aerial vehicle system |
| CN117853462A (en) * | 2024-01-11 | 2024-04-09 | 连云港市第二人民医院(连云港市临床肿瘤研究所) | Intra-articular pressure detection and information extraction method and system based on multi-mode imaging |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110427966A (en) | One kind rejecting error hiding feature point methods based on characteristic point local feature | |
| CN111243032B (en) | Full-automatic detection method for checkerboard corner points | |
| CN110443128B (en) | Finger vein identification method based on SURF feature point accurate matching | |
| Zhang et al. | Novel accurate and fast optic disc detection in retinal images with vessel distribution and directional characteristics | |
| CN108022228A (en) | Based on the matched colored eye fundus image joining method of SIFT conversion and Otsu | |
| CN110119741A (en) | A kind of card card image information recognition methods having powerful connections | |
| CN104933434A (en) | Image matching method combining length between perpendiculars (LBP) feature extraction method and surf feature extraction method | |
| Selvakumar et al. | The performance analysis of edge detection algorithms for image processing | |
| CN109949227A (en) | Image split-joint method, system and electronic equipment | |
| CN108961301B (en) | Chaetoceros image segmentation method based on unsupervised pixel-by-pixel classification | |
| CN108921813A (en) | Unmanned aerial vehicle detection bridge structure crack identification method based on machine vision | |
| CN109978848A (en) | Method based on hard exudate in multiple light courcess color constancy model inspection eye fundus image | |
| Lee et al. | Edge detection analysis | |
| CN109658391A (en) | A kind of radius of circle measurement method being fitted based on contour mergence and convex closure | |
| CN104123554A (en) | SIFT image characteristic extraction method based on MMTD | |
| Pavlović et al. | Edge detection parameter optimization based on the genetic algorithm for rail track detection | |
| CN106485252A (en) | Dot matrix target image Feature point recognition method is tested in image registration | |
| CN109977910B (en) | Rapid bill positioning method and system based on color line segments | |
| CN109712116B (en) | Fault identification method for power transmission line and accessories thereof | |
| CN106204648B (en) | A kind of method for tracking target and device rejected based on background | |
| Aktaş et al. | Edge-aware segmentation in satellite imagery: A case study of shoreline detection | |
| CN117078726A (en) | Different spectrum image registration method based on edge extraction | |
| CN115690107A (en) | High-throughput counting method, system and device for pod fruit grains based on image processing | |
| CN111667429A (en) | Target positioning and correcting method for inspection robot | |
| CN111640058A (en) | Image fusion processing method and device |
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 | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Yu Zhenjun Inventor after: Jia Shang Feng Inventor before: Yu Zhenjun Inventor before: Sun Lin Inventor before: Jia Shang Feng |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191108 |