CN103337068A - A multiple-subarea matching method constraint by a space relation - Google Patents
A multiple-subarea matching method constraint by a space relation Download PDFInfo
- Publication number
- CN103337068A CN103337068A CN2013102193926A CN201310219392A CN103337068A CN 103337068 A CN103337068 A CN 103337068A CN 2013102193926 A CN2013102193926 A CN 2013102193926A CN 201310219392 A CN201310219392 A CN 201310219392A CN 103337068 A CN103337068 A CN 103337068A
- Authority
- CN
- China
- Prior art keywords
- subtemplate
- matching
- subarea
- templates
- sub
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a multiple-subarea matching method constraint by a space relation. A plurality of subareas are selected in a matching template as sub-templates, and the sub-templates are separately and simultaneously matched with to-be-matched images. Space position relations among the sub-templates are utilized to realize the accurate matching of the images. The method comprises the following steps: the plurality of subareas are selected in the matching template respectively as the sub-templates; the space relations among the sub-templates are determined; the plurality of sub-templates keep the space position relations unchanged to form combined templates; matching is carried out by utilizing the mobile searching performed by the combined templates on the to-be-matched images, and a plurality of similarity values of the combined templates are obtained; the plurality of similarity values are compared, and a position where a maximum similarity value exists is regarded as an optimum matching position to complete the matching. The method of the invention utilizes the mutual space relations among the plurality of sub-templates to reach the requirements for accuracy and precision of matching. The object identifying performance of the method is substantially improved in real-time performance compared with the conventional large-template gray scale or contour matching algorithms.
Description
Technical field
The invention belongs to digital picture matching technique field, be specifically related to many subareas of a kind of image matching process.
Background technology
Along with science and technology especially computer technology rapid development, make the image processing techniques of directly from image, obtaining information obtain development at full speed, images match is one of very important technology during computer vision and image are handled.
Images match is exactly in a width of cloth or several unknown images, calculates the process of seeking the subgraph corresponding with known pattern by coupling.At present, image matching technology is applied in every field such as military affairs, industry, remote sensing, medical science and machine vision widely.
In the images match of reality is used, need to select the template of suitable size.But in coupling was calculated, if increase template figure, the increase that the calculated amount of coupling can be violent caused adapting to the occasion that real-time is had relatively high expectations, if reduce the size of template in order to reduce calculated amount, then can influence accuracy and the precision of coupling.In the application scenario that real-time is had relatively high expectations, accuracy and the precision of the coupling of having to sometimes reduce in order to guarantee real-time.
In nearest technology, the more similar approach that adopts template matches is also arranged, such as bright, people's such as field " infrared forward sight is studied a class special building target identification technology " (coming from Acta Astronautica's the 31st the 4th phase of volume of April in 2010) literary composition, characteristics at infrared/visible light multi-mode image coupling, computing method based on the gradient vector related coefficient have been proposed, lost the shortcoming of gradient direction information for gradient intensity related coefficient computing method, this method uses the gradient vector field to mate, avoided losing of directional information, adopt the similarity measure of large form coupling, matching performance is greatly improved.But in the occasion that real-time is had relatively high expectations, this method can be short of to some extent, can't take full advantage of the important information of template in short computing time, reaches the effect of quick calculating.
The method of existing raising matching speed is difficult to guarantee the better matching property energy, the accuracy and the precision that are coupling differ bigger with respect to the matching algorithm that does not raise speed, it is the less stable that matching speed is improved, can reduce more calculated amount in some cases, improve more matching speed, but almost can not reduce calculated amount in some cases, the computing velocity of the matching algorithm after namely raising speed and the matching algorithm that does not raise speed is similar substantially.
Summary of the invention
Above defective or improvement demand at prior art, the invention provides a kind of many subareas matching process based on the spatial relationship constraint, its purpose is by a large form is split as several less subtemplates, each subtemplate keeps the position relation the same with large form, solve thus under the situation that guarantees coupling accuracy and precision, reduce calculated amount and reach the technical matters that improves arithmetic speed.
As follows for realizing the concrete technical scheme that purpose of the present invention takes:
(1) chooses the subarea
Some zonules that will belong in the big zone are called the subarea, choose several subareas in template, and this subarea is also namely as subtemplate.
The principle that the subarea is selected generally be choose angle point, line feature obviously, have the zone that is different from other subareas or has certain discrimination, avoid the more zone of repeat pattern as far as possible.The size in each subarea should not be too little, otherwise precision and the accuracy to coupling has bigger influence when yardstick or angular error are arranged.Should note the spatial relationship constraint when subarea is selected, each subarea is selected not want adjacent too near, should make these subareas can comprise the large form most information, so also can reduce error, improves matching precision.
(2) calculate the subarea spatial relation
Spatial relation between the subarea calculates by template figure and gets.Computing method are as follows:
Earlier with any one subarea as benchmark, be true origin with any point in the benchmark subarea (for example can with upper left angle point or central point), by calculating the distance of this any point on other subarea respective point and this benchmark subarea, obtain the coordinate points in other subarea, thereby obtain the spatial relation in other each subareas and benchmark subarea.
Because what calculate is the relative position relation in each subarea, the position relation between the subarea is fixed, and therefore no matter chooses which subarea as benchmark, and the spatial relation in each sub-range is constant, to not influence of matching performance.
(3) coupling is calculated
Coupling namely refers to utilize the subarea in the reference map to search identical or the most close matching area in real-time figure to be matched.When mating calculating, keep each subarea spatial relation constant that each subtemplate is mobile to mate at image to be matched simultaneously.Matching degree is weighed by similarity, and matching process namely is the process of calculating similarity.
When calculating similarity, can adopt the similarity value in each subarea of independent calculating, and then similarity (degree of confidence) value in each subarea added up mutually, obtaining a similarity value sum, this similarity value is as the similarity degree of weighing the zone that above-mentioned a plurality of subareas constitute.
Calculation of similarity degree has the method for many maturations, and how explanation calculates the similarity value R in a subarea with regard to the normalized crosscorrelation method now.
Calculating such as the formula (1) of normalized crosscorrelation method (Normal cross-correlation is called for short NCC):
In the formula: (x y) is the similarity value, I(i to R, j) be that size is image to be matched for the search graph of W * H, T(i is that size is the subarea template of M * N j), wherein, (i j) is any pixel, M, N, W, H is positive integer, represent length and the width of search graph respectively, and the length of subarea template and width.(x is that template covers the coordinate of subgraph any point (for example top left corner apex) in search graph y).
Certainly, when calculating the similarity value, also can adopt and regard each subarea as a large form calculating, calculate an overall similarity value.
(4) obtain optimal match point
In the region of search, compare these similarity values, get the position of similarity maximum as best match position.
Method of the present invention is by being split as several less subtemplates with a large form, each subtemplate keeps the position relation the same with large form, and only comprise information important in the large form in each subtemplate, ignore the little information of reusability in the large form, fully guaranteeing under the situation of coupling accuracy and precision thereby reach, applicable to many occasions, guaranteeing bigger raising matching speed under coupling accuracy and the precision prerequisite, reduce calculated amount largely, reach the purpose that improves arithmetic speed.
Many subareas matching process by spatial relationship constraint of the present invention carries out images match, both taken full advantage of the important information among the large form figure, neglect in the large form the little information of reusability in computation process, reduced calculated amount, requirement of real time, also can utilize the mutual spatial relationship between each subarea, reach matched accuracy and accuracy requirement.A large amount of test findings show that with respect to traditional large form gray scale or outline algorithm, real-time performance is greatly improved this method in target identification performance.
Description of drawings
Fig. 1 is many subareas matching process schematic flow sheet of the spatial relationship constraint of the embodiment of the invention.
Fig. 2 is the scene graph of the embodiment of the invention when specifically using.
Fig. 3 is the target large form figure that the embodiment of the invention is chosen.
Fig. 4 is many subareas synoptic diagram that the embodiment of the invention is chosen.
Fig. 5 is the synoptic diagram of the mutual alignment relation in each subarea of the embodiment of the invention.
Fig. 6 is the embodiment of the invention each subarea coordinate on figure to be matched when coupling is calculated.
Embodiment
In order to make purpose of the present invention, technical scheme and advantage clearer, below in conjunction with drawings and Examples, the present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explaining the present invention, and be not used in restriction the present invention.
As shown in Figure 1, the concrete steps based on many subareas matching process of spatial relationship constraint of present embodiment are as follows:
(1) chooses the subarea
Scene graph and target are mated the selected template figure of target as shown in Figure 3 as shown in Figure 2.The subarea selection principle is introduced in characteristics and front according to template figure, chosen the subarea that a plurality of (for example 5) have certain discrimination at template figure, mainly comprise angle point, line feature in these zones, have the special shape or other contents with certain discrimination that are different from other subareas, the zone of choosing as shown in Figure 4.
(2) obtain the spatial relation in each subarea
After having chosen the subarea, also need to obtain the spatial relation in each subarea.
Can get any one subarea as benchmark, be true origin with certain point (being preferably upper left angle point or central point) in the benchmark subarea, by calculating the relative distance in other subarea and benchmark subarea, the calculating of this distance, can read their relative coordinate by they relative positions on large form, obtain the coordinate figure in other subarea.
Here the upper left angle point with the benchmark subarea is example as true origin, other subareas on template figure in the spatial relation in benchmark subarea as shown in Figure 5.
If the image of high-level different visual angles object observing, because the existence of perspective transform, change has all taken place in the size of its image, shape, and following view picture is different with the angle that preceding view picture is down taken, both exist bigger disparity, bring difficulty to registration.In order to reduce visual angle, high-altitude influence, before mating, at first should be with each subtemplate from apparent direction perspective transform down to preceding apparent direction, and then utilize the subtemplate after the conversion to mate.
(3) coupling is calculated
When mating calculating, keep each subarea spatial relation to mate, namely each subarea keeps its position relationship and mates at the enterprising line search of image to be matched simultaneously together.
In order to keep the spatial relation in each subarea, can determine the position in benchmark subarea earlier, by the mutual alignment relation in other subareas and benchmark subarea, calculate the position of other subareas on figure to be matched then, mate calculating again.
For example, the coordinate of benchmark subarea on image to be matched is (70,221), and then the coordinate of other subarea on figure to be matched calculates by the relation of the mutual alignment between them, as shown in Figure 6.
After having determined the matched position of each subarea on figure to be matched, just can calculation template and the similarity of figure to be matched.
When carrying out measuring similarity calculating, calculate the similarity in each subarea earlier separately, and then the similarity in each subarea is added up mutually, obtain last result as shown in Equation 2.
c＝c
1+c
2+…+c
n（2）
Wherein c is the similarity that position calculation come out of template figure on real-time figure, c
1, c
2C
nThe similarity of representing each subarea, n represents the quantity in subarea.
Certainly, also can adopt each subarea is calculated as the various piece of a large overall template, namely when calculating, regard each subarea combination as a large form, directly calculate an overall similarity value c, this kind regarded each subarea as calculating that integral body carries out and separate computations subarea similarity, and the method for addition is similar again, mates the advantage that also has on the speed with respect to the full detail of direct use large form.This kind utilizes the method for many subareas coupling of spatial relationship constraint, the similarity that each subarea that combines is obtained adds up, take full advantage of the key character information that comprises among the large form figure, got rid of the characteristic information of redundancy in the large form, reduced calculated amount, so have more advantage in the coupling real-time.
(4) obtain optimal match point
At last, in the region of search, compare these similarity values, obtain the position of similarity maximum, as best match position, finish coupling.
Those skilled in the art will readily understand; the above only is preferred embodiment of the present invention; not in order to limiting the present invention, all any modifications of doing within the spirit and principles in the present invention, be equal to and replace and improvement etc., all should be included within protection scope of the present invention.
Claims (6)
1. many subareas matching process based on spatial relationship constraint, it mates respectively as subtemplate and image to be matched simultaneously by choose a plurality of subregions in matching template, utilize the spatial relation of each subtemplate to realize that image accurately mates, it is characterized in that this method specifically comprises:
Choose several subregions at matching template, respectively as subtemplate;
Determine the spatial relation between each subtemplate;
Each subtemplate keeps the indeformable one-tenth gang form of its spatial relation, utilizes the mobile search on image to be matched of this gang form to obtain the similarity value of a plurality of these gang forms to mate;
More above-mentioned a plurality of similarity value, and with the position of similarity value maximum wherein as best match position, finish coupling.
2. the many subareas matching process based on the spatial relationship constraint according to claim 1 is characterized in that the similarity value of described gang form can obtain by the similarity value addition of each subtemplate in this gang form.
3. the many subareas matching process based on the spatial relationship constraint according to claim 1 and 2 is characterized in that, the spatial relation of described each subtemplate is determined by the distance of corresponding point in each subtemplate.
4. the many subareas matching process based on the spatial relationship constraint according to claim 3 is characterized in that, the spatial relation of described each subtemplate is following to be determined:
As benchmark, and be true origin with any point in this benchmark subtemplate with any subtemplate, calculate the point of other subtemplate corresponding position and the distance of this this any point, can obtain the spatial relation between other subtemplate and the benchmark subtemplate.
5. the many subareas matching process based on the spatial relationship constraint according to claim 4 is characterized in that, any point in the described benchmark subtemplate can be gone up other arbitrfary points for central point, frontier point or the zone in subtemplate zone.
6. according to each described many subareas matching process based on the spatial relationship constraint among the claim 1-5, it is characterized in that described all subregion is separate, each does not overlap.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310219392.6A CN103337068B (en) | 2013-06-04 | 2013-06-04 | The multiple subarea matching process of spatial relation constraint |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310219392.6A CN103337068B (en) | 2013-06-04 | 2013-06-04 | The multiple subarea matching process of spatial relation constraint |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103337068A true CN103337068A (en) | 2013-10-02 |
| CN103337068B CN103337068B (en) | 2015-09-09 |
Family
ID=49245216
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310219392.6A Expired - Fee Related CN103337068B (en) | 2013-06-04 | 2013-06-04 | The multiple subarea matching process of spatial relation constraint |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103337068B (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103761730A (en) * | 2013-12-31 | 2014-04-30 | 华中科技大学 | Knowledge-constrained road vehicle target image pneumatic optical effect correcting method |
| CN105260740A (en) * | 2015-09-23 | 2016-01-20 | 广州视源电子科技股份有限公司 | Element recognition method and apparatus |
| CN106503737A (en) * | 2016-10-20 | 2017-03-15 | 广州视源电子科技股份有限公司 | A kind of electronic units fix method and device |
| CN106996911A (en) * | 2016-01-22 | 2017-08-01 | 名硕电脑(苏州)有限公司 | The localization method of two-dimensional detection paste solder printing |
| CN107590517A (en) * | 2017-09-19 | 2018-01-16 | 安徽大学 | Image similarity measurement method and image registration method based on shape information |
| CN107734268A (en) * | 2017-09-18 | 2018-02-23 | 北京航空航天大学 | A kind of structure-preserved wide baseline video joining method |
| CN108010068A (en) * | 2017-11-29 | 2018-05-08 | 中国人民解放军火箭军工程大学 | Ground time critical target recognition methods based on gradient direction characteristic point pair |
| CN109410175A (en) * | 2018-09-26 | 2019-03-01 | 北京航天自动控制研究所 | SAR radar imagery quality quick automatic evaluation method based on multiple subarea images match |
| CN110288040A (en) * | 2019-06-30 | 2019-09-27 | 北京华融塑胶有限公司 | A kind of similar evaluation method of image based on validating topology and equipment |
| CN111915898A (en) * | 2020-07-24 | 2020-11-10 | 杭州金通科技集团股份有限公司 | Parking monitoring AI electronic post house |
| CN112634140A (en) * | 2021-03-08 | 2021-04-09 | 广州松合智能科技有限公司 | High-precision full-size visual image acquisition system and method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102194131A (en) * | 2011-06-01 | 2011-09-21 | 华南理工大学 | Fast human face recognition method based on geometric proportion characteristic of five sense organs |
| CN102663754A (en) * | 2012-04-17 | 2012-09-12 | 北京博研新创数码科技有限公司 | Image matching calculation method based on regional Gaussian weighting |
-
2013
- 2013-06-04 CN CN201310219392.6A patent/CN103337068B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102194131A (en) * | 2011-06-01 | 2011-09-21 | 华南理工大学 | Fast human face recognition method based on geometric proportion characteristic of five sense organs |
| CN102663754A (en) * | 2012-04-17 | 2012-09-12 | 北京博研新创数码科技有限公司 | Image matching calculation method based on regional Gaussian weighting |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103761730B (en) * | 2013-12-31 | 2016-04-13 | 华中科技大学 | The road vehicle target image aero-optical effect bearing calibration of knowledge constraints |
| CN103761730A (en) * | 2013-12-31 | 2014-04-30 | 华中科技大学 | Knowledge-constrained road vehicle target image pneumatic optical effect correcting method |
| CN105260740A (en) * | 2015-09-23 | 2016-01-20 | 广州视源电子科技股份有限公司 | Element recognition method and apparatus |
| WO2017050083A1 (en) * | 2015-09-23 | 2017-03-30 | 广州视源电子科技股份有限公司 | Element identification method and device |
| CN106996911A (en) * | 2016-01-22 | 2017-08-01 | 名硕电脑(苏州)有限公司 | The localization method of two-dimensional detection paste solder printing |
| WO2018072334A1 (en) * | 2016-10-20 | 2018-04-26 | 广州视源电子科技股份有限公司 | Method and device for positioning electronic component |
| CN106503737A (en) * | 2016-10-20 | 2017-03-15 | 广州视源电子科技股份有限公司 | A kind of electronic units fix method and device |
| CN106503737B (en) * | 2016-10-20 | 2019-03-05 | 广州视源电子科技股份有限公司 | A kind of electronic units fix method and device |
| CN107734268A (en) * | 2017-09-18 | 2018-02-23 | 北京航空航天大学 | A kind of structure-preserved wide baseline video joining method |
| CN107590517A (en) * | 2017-09-19 | 2018-01-16 | 安徽大学 | Image similarity measurement method and image registration method based on shape information |
| CN107590517B (en) * | 2017-09-19 | 2020-07-10 | 安徽大学 | Image similarity measurement method and image registration method based on shape information |
| CN108010068A (en) * | 2017-11-29 | 2018-05-08 | 中国人民解放军火箭军工程大学 | Ground time critical target recognition methods based on gradient direction characteristic point pair |
| CN109410175B (en) * | 2018-09-26 | 2020-07-14 | 北京航天自动控制研究所 | SAR radar imaging quality rapid automatic evaluation method based on multi-subregion image matching |
| CN109410175A (en) * | 2018-09-26 | 2019-03-01 | 北京航天自动控制研究所 | SAR radar imagery quality quick automatic evaluation method based on multiple subarea images match |
| CN110288040A (en) * | 2019-06-30 | 2019-09-27 | 北京华融塑胶有限公司 | A kind of similar evaluation method of image based on validating topology and equipment |
| CN110288040B (en) * | 2019-06-30 | 2022-02-11 | 北京华融塑胶有限公司 | Image similarity judging method and device based on topology verification |
| CN111915898A (en) * | 2020-07-24 | 2020-11-10 | 杭州金通科技集团股份有限公司 | Parking monitoring AI electronic post house |
| CN112634140A (en) * | 2021-03-08 | 2021-04-09 | 广州松合智能科技有限公司 | High-precision full-size visual image acquisition system and method |
| CN112634140B (en) * | 2021-03-08 | 2021-08-13 | 广州松合智能科技有限公司 | High-precision full-size visual image acquisition system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103337068B (en) | 2015-09-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103337068A (en) | A multiple-subarea matching method constraint by a space relation | |
| US10867189B2 (en) | Systems and methods for lane-marker detection | |
| CN103714548B (en) | Infrared image and visible image registration method based on visual attention | |
| CN104700414B (en) | A kind of road ahead pedestrian's fast ranging method based on vehicle-mounted binocular camera | |
| US9349189B2 (en) | Occlusion resistant image template matching using distance transform | |
| CN103411609B (en) | A kind of aircraft return route planing method based on online composition | |
| CN103604417B (en) | The multi-view images bi-directional matching strategy that object space is information constrained | |
| CN105184830B (en) | A kind of symmetrical shaft detection localization method of symmetric graph picture | |
| CN106052646A (en) | Information processing apparatus and information processing method | |
| CN103700099A (en) | Rotation and dimension unchanged wide baseline stereo matching method | |
| CN104268880A (en) | Depth information obtaining method based on combination of features and region matching | |
| CN107909018B (en) | Stable multi-mode remote sensing image matching method and system | |
| CN103093459A (en) | Assisting image matching method by means of airborne lidar point cloud data | |
| CN103247029A (en) | Geometric registration method for hyperspectral image generated by spliced detectors | |
| Nilosek et al. | Assessing geoaccuracy of structure from motion point clouds from long-range image collections | |
| CN105387847A (en) | Non-contact measurement method, measurement equipment and measurement system thereof | |
| Li et al. | Seamless positioning and navigation by using geo-referenced images and multi-sensor data | |
| CN104318566A (en) | Novel multi-image plumb line track matching method capable of returning multiple elevation values | |
| Cheng et al. | Positioning and navigation of mobile robot with asynchronous fusion of binocular vision system and inertial navigation system | |
| Wang et al. | Line matching of wide baseline images in an affine projection space | |
| CN104166977A (en) | Image matching similarity measuring method and image matching method thereof | |
| Wan et al. | The P2L method of mismatch detection for push broom high-resolution satellite images | |
| WO2015027649A1 (en) | Vehicle detection method using multi-scale model | |
| CN103473756B (en) | Autoregistration merges the method for Aerial Images and close shot image | |
| US20230104937A1 (en) | Absolute scale depth calculation device, absolute scale depth calculation method, and computer program product |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20150909 Termination date: 20200604 |