CN103337068B - The multiple subarea matching process of spatial relation constraint - Google Patents
The multiple subarea matching process of spatial relation constraint Download PDFInfo
- Publication number
- CN103337068B CN103337068B CN201310219392.6A CN201310219392A CN103337068B CN 103337068 B CN103337068 B CN 103337068B CN 201310219392 A CN201310219392 A CN 201310219392A CN 103337068 B CN103337068 B CN 103337068B
- Authority
- CN
- China
- Prior art keywords
- subtemplate
- subarea
- spatial relation
- matching
- similarity value
- 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.)
- Expired - Fee Related
Links
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a kind of multiple subarea matching process based on spatial relation constraint, it mates with image to be matched as subtemplate respectively by choosing multiple subregion in matching template simultaneously, the spatial relation of each subtemplate is utilized to realize accurate matching of image, it comprises: choose several subregions, respectively as subtemplate at matching template; Determine the spatial relation between each subtemplate; Each subtemplate keeps the indeformable one-tenth gang form of its spatial relation, utilizes gang form mobile search on image to be matched to mate, to obtain the Similarity value of this gang form multiple; More above-mentioned multiple Similarity value, and using the maximum position of wherein Similarity value as best match position, complete coupling.Method of the present invention utilizes the mutual spatial relationship between each subarea, reaches accuracy and the accuracy requirement of coupling, and the method is in target identification performance relative to traditional large form gray scale or outline algorithm, and real-time performance is greatly improved.
Description
Technical field
The invention belongs to Image-matching technical field, be specifically related to a kind of image multiple subarea matching process.
Background technology
Along with developing rapidly of science and technology especially computer technology, make directly from image the image processing techniques of obtaining information obtain development at full speed, images match is one of very important technology in computer vision and image procossing.
Images match is exactly in the image an of width or several the unknowns, is found the process of the subgraph corresponding with known pattern by matching primitives.At present, image matching technology is applied in the every field such as military affairs, industry, remote sensing, medical science and machine vision widely.
In the images match application of reality, need the template selecting suitable size.But in matching primitives, if increase Prototype drawing, the increase that the calculated amount of coupling can be violent, causing adapting to the higher occasion of requirement of real-time, if reduce the size of template to reduce calculated amount, then can affect accuracy and the precision of coupling.In the application scenario higher to requirement of real-time, have to sometimes to ensure that real-time reduces accuracy and the precision of coupling.
In nearest technology, also the more similar approach that have employed template matches is had, such as bright, " infrared forward sight is studied a class special building target identification technology " (coming from Acta Astronautica's volume the 4th phase April the 31st in 2010) one literary composition of the people such as field, for the feature of infrared/visible ray multi-mode image coupling, propose the computing method based on gradient vector related coefficient, gradient intensity Calculation of correlation factor method be lost to the shortcoming of Gradient direction information, the method uses gradient vector field to mate, avoid the loss of directional information, adopt the similarity measure of large form coupling, matching performance is greatly improved.But in the occasion that requirement of real-time is higher, the method can be short of to some extent, the important information of template cannot be made full use of within shorter computing time, reach the effect calculated fast.
The method of existing raising matching speed is difficult to ensure good matching performance, the accuracy and the precision that are coupling are larger relative to the matching algorithm difference do not raised speed, it is the less stable that matching speed is improved, more calculated amount can be reduced in some cases, improve more matching speed, but almost can not reduce calculated amount in some cases, the matching algorithm namely after speed-raising and the computing velocity of matching algorithm do not raised speed substantially similar.
Summary of the invention
For above defect or the Improvement requirement of prior art, the invention provides a kind of multiple subarea matching process based on spatial relation constraint, its object is to by a large form is split as several less subtemplates, each subtemplate keeps the position relationship the same with large form, solving when ensureing coupling accuracy and precision thus, reducing calculated amount and reaching the technical matters improving arithmetic speed.
As follows for realizing the concrete technical scheme that object of the present invention takes:
(1) subarea is chosen
The some zonules belonged in large regions are called subarea, template is chosen several subareas, this subarea is also namely as subtemplate.
The principle that subarea is selected be generally choose angle point, line features obviously, there is the region being different from other subareas or having certain discrimination, avoid the region that repeat pattern is more as far as possible.The size in each subarea should not be too little, otherwise have larger impact when having yardstick or angular error to the precision of coupling and accuracy.Should note spatial relation constraint when subarea is selected, each subarea is selected not adjacent too near, these subareas should be made can to comprise large form most information, so also can reduce error, improve matching precision.
(2) subarea spatial relation is calculated
Spatial relation between subarea is calculated by Prototype drawing and gets.Computing method are as follows:
First using any one subarea as benchmark, with any point (such as can use upper left angle point or central point) in benchmark subarea for true origin, by calculating the distance of this any point on other subarea respective point and this benchmark subarea, obtain the coordinate points in other subarea, thus obtain the spatial relation in other each subareas and benchmark subarea.
Be the relative position relation in each subarea due to what calculate, the position relationship between subarea is fixing, and therefore no matter choose which subarea as benchmark, the spatial relation in each sub-range is constant, does not affect matching performance.
(3) matching primitives
Namely coupling refers to, in real-time figure to be matched, utilize the subarea in reference map to search identical or the most close matching area.When carrying out matching primitives, keep each subarea spatial relation constant that each subtemplate is simultaneously mobile to mate on image to be matched.Matching degree is weighed by similarity, and namely matching process is the process calculating similarity.
When calculating similarity, the Similarity value calculating separately each subarea can be adopted, and then similarity (degree of confidence) value in each subarea is added up mutually, obtain a Similarity value sum, this Similarity value is as the similarity degree weighing the region that above-mentioned multiple subarea is formed.
The calculating of similarity has the method for many maturations, and the Similarity value R how calculating a subarea is described with regard to normalized crosscorrelation method now.
The calculating of normalized crosscorrelation method (Normal cross-correlation is called for short NCC) is as formula (1):
In formula: R (x, y) is Similarity value, I(i, j) be size be the search graph of W × H and image to be matched, T(i, j) the subarea template of to be size be M × N, wherein, (i, j) is any pixel, M, N, W, H are positive integer, represent length and the width of search graph respectively, and the length of subarea template and width.(x, y) be template cover subgraph any point (such as top left corner apex) coordinate in search graph.
Certainly, when calculating Similarity value, also can adopt and regarding each subarea as a large form calculating, calculating a global similarity angle value.
(4) optimal match point is obtained
In region of search, compare these Similarity value, get the maximum position of similarity as best match position.
Method of the present invention is by being split as several less subtemplates by a large form, each subtemplate keeps the position relationship the same with large form, and only comprise information important in large form in each subtemplate, ignore the information that reusability in large form is little, thus reach when fully ensureing coupling accuracy and precision, be applicable to many occasions, larger raising matching speed under guarantee coupling accuracy and precision prerequisite, reduce calculated amount largely, reach the object improving arithmetic speed.
Images match is carried out by the multiple subarea matching process of spatial relation constraint of the present invention, both the important information in large form figure had been taken full advantage of, neglect the information that in large form, reusability is little in computation process, decrease calculated amount, requirement of real time, also the mutual spatial relationship between each subarea be can utilize, accuracy and the accuracy requirement of coupling reached.Many experimental results shows, the method is in target identification performance relative to traditional large form gray scale or outline algorithm, and real-time performance is greatly improved.
Accompanying drawing explanation
Fig. 1 is the multiple subarea matching process schematic flow sheet of the spatial relation constraint of the embodiment of the present invention.
Scene graph when Fig. 2 is embodiment of the present invention embody rule.
Fig. 3 is the target large form figure that the embodiment of the present invention is chosen.
Fig. 4 is the multiple subarea schematic diagram that the embodiment of the present invention is chosen.
Fig. 5 is the schematic diagram of the mutual alignment relation in each subarea of the embodiment of the present invention.
Fig. 6 is the embodiment of the present invention each subarea coordinate on figure to be matched when matching primitives.
Embodiment
In order to make object of the present invention, technical scheme and advantage clearly understand, 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 explain the present invention, be not intended to limit the present invention.
As shown in Figure 1, the concrete steps of the multiple subarea matching process based on spatial relation constraint of the present embodiment are as follows:
(1) subarea is chosen
As shown in Figure 2, the Prototype drawing of coupling selected by target as shown in Figure 3 for scene graph and target.According to the feature of Prototype drawing with introduce subarea selection principle above, Prototype drawing have chosen the subarea that multiple (such as 5) have certain discrimination, mainly comprise angle point, line features in these regions, have the special shape that is different from other subareas or other have the content of certain discrimination, the region chosen as shown in Figure 4.
(2) spatial relation in each subarea is obtained
After have chosen subarea, also need the spatial relation obtaining each subarea.
Any one subarea can be got as benchmark, with certain point (being preferably upper left angle point or central point) in benchmark subarea for true origin, by calculating the relative distance in other subarea and benchmark subarea, the calculating of this distance, their relative coordinate can be read by their relative positions on large form, obtain the coordinate figure in other subarea.
Here for the upper left angle point in benchmark subarea as true origin, other subareas on Prototype drawing in the spatial relation in benchmark subarea as shown in Figure 5.
If the image of high-level different visual angles object observing, due to the existence of perspective transform, size, the shape of its image all there occurs change, and lower view picture is different with front lower angle of looking image taking, both also exist larger disparity, bring difficulty to registration.In order to reduce high-altitude angle effects, before mating, first should by each subtemplate from lower apparent direction perspective transform to front apparent direction, and then the subtemplate after conversion be utilized to mate.
(3) matching primitives
When carrying out matching primitives, 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 together simultaneously.
In order to keep the spatial relation in each subarea, first can determine the position in benchmark subarea, then passing through the mutual alignment relation in other subareas and benchmark subarea, calculating the position of other subareas on figure to be matched, then carry out matching primitives.
Such as, the coordinate of benchmark subarea on image to be matched is (70,221), then other the coordinate of subarea on figure to be matched, is calculated, as shown in Figure 6 by the mutual alignment relation between them.
After determining the matched position of each subarea on figure to be matched, just can the similarity of calculation template and figure to be matched.
When carrying out measuring similarity calculating, first calculating separately the similarity in each subarea, and then the similarity in each subarea is added up mutually, obtaining last result as shown in Equation 2.
c＝c
1+c
2+…+c
n（2）
Wherein c is the position calculation similarity out of Prototype drawing on real-time figure, c
1, c
2c
nrepresent the similarity in each subarea, n represents the quantity in subarea.
Certainly, also can adopt and the various piece of each subarea as a large overall template is calculated, namely the combination of each subarea is regarded as a large form when calculating, directly calculate a global similarity angle value c, it is similar that this kind regards each subarea the method that calculating that entirety carries out and separate computations subarea similarity be added again as, carries out mating the advantage also had in speed relative to directly using the full detail of large form.The method that this kind utilizes the multiple subarea of spatial relation constraint to mate, the similarity that each subarea combined obtains is added up, make full use of the key character information comprised in large form figure, eliminate the characteristic information of redundancy in large form, decrease calculated amount, so have more advantage in coupling real-time.
(4) optimal match point is obtained
Finally, in region of search, compare these Similarity value, obtain the position that similarity is maximum, as best match position, complete coupling.
Those skilled in the art will readily understand; the foregoing is only preferred embodiment of the present invention; not in order to limit the present invention, all any amendments done within the spirit and principles in the present invention, equivalent replacement and improvement etc., all should be included within protection scope of the present invention.
Claims (3)
1. the multiple subarea matching process based on spatial relation constraint, it mates with image to be matched as subtemplate respectively by choosing multiple subregion in matching template simultaneously, the spatial relation of each subtemplate is utilized to realize accurate matching of image, it is characterized in that, the method specifically comprises:
Choose several subregions at matching template, respectively as subtemplate, described all subregion is separate, does not respectively overlap;
Determine the spatial relation between each subtemplate, wherein the spatial relation of each subtemplate is determined by the distance of corresponding point in each subtemplate, namely using any subtemplate as benchmark, and with any point in this benchmark subtemplate for true origin, calculate the point of other subtemplate corresponding position and the distance of this this any point, the spatial relation between other subtemplate and benchmark subtemplate can be obtained;
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 mate, to obtain the Similarity value of this gang form multiple;
More above-mentioned multiple Similarity value, and using the maximum position of wherein Similarity value as best match position, complete coupling.
2. the multiple subarea matching process based on spatial relation constraint according to claim 1, is characterized in that, the Similarity value of described gang form is added by the Similarity value of each subtemplate in this gang form and obtains.
3. the multiple subarea matching process based on spatial relation constraint according to claim 1, is characterized in that, any point in described benchmark subtemplate can be other arbitrfary points on the central point in subtemplate region, frontier point or region.
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 CN103337068A (en) | 2013-10-02 |
| CN103337068B true 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) |
Families Citing this family (11)
| 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 |
| CN105260740B (en) * | 2015-09-23 | 2019-03-29 | 广州视源电子科技股份有限公司 | Element identification method and device |
| CN106996911B (en) * | 2016-01-22 | 2019-08-20 | 名硕电脑(苏州)有限公司 | The localization method of two-dimensional detection paste solder printing |
| CN106503737B (en) * | 2016-10-20 | 2019-03-05 | 广州视源电子科技股份有限公司 | Electronic component positioning 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 |
| 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 |
| 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 |
| CN111915898B (en) * | 2020-07-24 | 2022-07-08 | 杭州金通科技集团股份有限公司 | Parking monitoring AI electronic post house |
| 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 |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102194131B (en) * | 2011-06-01 | 2013-04-10 | 华南理工大学 | Fast human face recognition method based on geometric proportion characteristic of five sense organs |
| CN102663754B (en) * | 2012-04-17 | 2014-12-10 | 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
Also Published As
| Publication number | Publication date |
|---|---|
| CN103337068A (en) | 2013-10-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103337068B (en) | The multiple subarea matching process of spatial relation constraint | |
| Jeong et al. | Complex urban dataset with multi-level sensors from highly diverse urban environments | |
| US8422825B1 (en) | Method and system for geometry extraction, 3D visualization and analysis using arbitrary oblique imagery | |
| US9330504B2 (en) | 3D building model construction tools | |
| US20150003741A1 (en) | Occlusion Resistant Image Template Matching Using Distance Transform | |
| Qu et al. | Landmark based localization in urban environment | |
| CN103604417B (en) | The multi-view images bi-directional matching strategy that object space is information constrained | |
| US10438362B2 (en) | Method and apparatus for homography estimation | |
| JP6334927B2 (en) | Additional information display device and additional information display program | |
| US11367213B2 (en) | Method and apparatus with location estimation | |
| Shi et al. | Fusion of a panoramic camera and 2D laser scanner data for constrained bundle adjustment in GPS-denied environments | |
| Remondino et al. | Surface reconstruction algorithms for detailed close-range object modeling | |
| Tao et al. | Automated processing of mobile mapping image sequences | |
| CN104318566A (en) | Novel multi-image plumb line track matching method capable of returning multiple elevation values | |
| TW202022804A (en) | Method and system for road image reconstruction and vehicle positioning | |
| Wang et al. | Line matching of wide baseline images in an affine projection space | |
| Müller et al. | Robust image registration for fusion | |
| Wan et al. | The P2L method of mismatch detection for push broom high-resolution satellite images | |
| Tsai et al. | Vision-based roadway geometry computation | |
| Yuan et al. | Estimation of vehicle pose and position with monocular camera at urban road intersections | |
| CN115830116A (en) | Robust visual odometer method | |
| Wang et al. | Positionit: an image-based remote target localization system on smartphones | |
| Mo et al. | Cross-based dense depth estimation by fusing stereo vision with measured sparse depth | |
| Lu | New efficient vanishing point detection from a single road image based on intrinsic line orientation and color texture properties | |
| Ghosh et al. | On localizing a camera from a single image |
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 |
Granted publication date: 20150909 Termination date: 20200604 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |