CN106355613B - The method for automatically extracting cross centre of figure based on least square fitting iteration - Google Patents
The method for automatically extracting cross centre of figure based on least square fitting iteration Download PDFInfo
- Publication number
- CN106355613B CN106355613B CN201610786591.9A CN201610786591A CN106355613B CN 106355613 B CN106355613 B CN 106355613B CN 201610786591 A CN201610786591 A CN 201610786591A CN 106355613 B CN106355613 B CN 106355613B
- Authority
- CN
- China
- Prior art keywords
- iteration
- dark
- initial coordinate
- fitting
- pixel
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 15
- 230000011218 segmentation Effects 0.000 claims abstract description 9
- 239000000284 extract Substances 0.000 claims abstract description 6
- 239000013078 crystal Substances 0.000 abstract description 16
- 230000003287 optical effect Effects 0.000 description 6
- 230000000694 effects Effects 0.000 description 4
- 230000005684 electric field Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000002441 X-ray diffraction Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000010287 polarization Effects 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
Landscapes
- Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of methods for automatically extracting cross centre of figure based on least square fitting iteration, belong to Digital image technology field, the dark right-angled intersection region solved in crystal conoscopic interference figure in the prior art is larger, it is difficult to the problem of accurately extracting cross centre of figure position.Steps are as follows by the present invention: obtaining digital picture；Digital picture is subjected to Threshold segmentation or edge extracting, extracts dark cross region；The initial coordinate in the dark cross region of iteration is set；Using initial coordinate as origin, by dark cross region segmentation at horizontal and vertical two parts, least square linear fit is weighted to horizontal and vertical two parts, obtains the two intersection point；It finds intersection at a distance from initial coordinate, and by intersecting point coordinate assignment in initial coordinate；Will be apart from as iteration convergence condition, the iteration ends if meeting the condition of convergence export coordinate centered on initial coordinate, otherwise iteration.The present invention is used to extract the center of cross figure.
Description
Technical field
A method of cross centre of figure being automatically extracted based on least square fitting iteration, for extracting cross figure
Center belongs to Digital image technology field.
Background technique
When light is propagated in crystal, optical frequency electric field and extra electric field can cause the nonlinear polarization of crystal, this phenomenon
Referred to as electrooptic effect.Crystal with electrooptic effect is known as electro-optic crystal.Using the electrooptic effect of crystal, pass through extra electric field
Variation achievees the purpose that photoelectric effect is mutually converted or modulated, and crystal can be made into the laser devices such as electrooptical switching, electro-optic deflector.
The optical axis direction of electrooptical switching crystal is direct to the dead axle of optical axis direction perpendicular to electrooptical switching crystal light passing surface
Affect the performance indicator of electrooptical switching.Electro-optic crystal often carries out optical axis direction with x-ray diffraction method and conoscopic interference method
Dead axle.When using conoscopic interference measurement electro-optic crystal optical axis direction, due to the birefringence of crystal, formed on diffuser screen
Dark cross searching position and optical axis of crystal side with the conoscopic interference figure that concentric loop and dark cross form, in conoscopic interference figure
To related.Accurately extract to the dark cross searching position of conoscopic interference figure is the key that guarantee high measurement accuracy.
Dark cross in crystal conoscopic interference figure is more coarse, and right-angled intersection region is big, utilizes the side such as image reform position
Method is more sensitive to factors such as light intensity uniformity, noises, it is difficult to accurately extract center, thus need to establish one kind and be suitable for
The accurate extracting method of big region cross centre of figure.
Summary of the invention
The present invention provides one kind in view of the above shortcomings and automatically extracts cross figure based on least square fitting iteration
The method at center, the dark right-angled intersection region that solution is solved in the prior art in crystal conoscopic interference figure in the prior art is larger,
It is difficult to the problem of accurately extracting cross centre of figure position.
To achieve the goals above, the technical solution adopted by the present invention are as follows:
A method of cross centre of figure being automatically extracted based on least square fitting iteration, which is characterized in that step is such as
Under:
(1) digital picture is obtained；
(2) digital picture is subjected to Threshold segmentation, according to dark cross intensity profile, be arranged suitable gray value to image into
Row binaryzation extracts the dark cross region in digital picture；
(3) initial coordinate (x in the dark cross region of iteration is set0,y0), that is, the geometric center position of digital picture is set
As initial coordinate (x0,y0)；
(4) with initial coordinate (x0,y0) be origin, by dark cross region segmentation at horizontal and vertical two parts, to laterally and
Longitudinal two parts are weighted least square linear fit, obtain the two intersection point (x1,y1)；
(5) (x is found intersection1,y1) and initial coordinate (x0,y0) distanceAnd by (x1,
y1) assignment is in initial coordinate (x0,y0)；
(6) will distance d as iteration convergence condition, the iteration ends if meeting the condition of convergence, export initial coordinate (x0,
y0) centered on coordinate, otherwise go to step (4).
Further, in the step (4), least square linear fit is weighted to horizontal and vertical two parts, is obtained
The two intersection point (x1,y1) specific steps are as follows:
(41) y=a is denoted as to the fitting a straight line of lateral part1x+b1, then lateral least square fitting formula is as follows:
In formula, subscript i represents different pixels, xi,yiIt is the transverse and longitudinal of each pixel in image dark ten block lateral part domain respectively
Coordinate, ciIt is the corresponding weighting coefficient of each pixelIiFor the gray value of each pixel, m is the dark cross region of image
The total pixel number of lateral part；
(42) x=a is denoted as to the fitting a straight line of longitudinal portion2y+b2, then longitudinal least square fitting formula is as follows:
In formula, subscript i represents different pixels, xi,yiIt is the transverse and longitudinal of the dark each pixel in ten blocks longitudinal portion domain of image respectively
Coordinate, ciIt is the corresponding weighting coefficient of each pixelIiFor the gray value of each pixel, m is the dark cross region of image
The total pixel number of longitudinal portion；
(43) according to lateral part and longitudinal portion fitting a straight line equation, the two intersection point (x is acquired1,y1)；
A in formula1,b1It is the slope and intercept of lateral fitting a straight line, a respectively2,b2It is the slope of longitudinal fitting a straight line respectively
And intercept.
Compared with the prior art, the advantages of the present invention are as follows:
According to the spatial distribution of cross figure, figure is divided into horizontal and vertical region, passes through weighted least-squares method point
The fitting a straight line for not fitting horizontal and vertical region, have the advantages that it is anti-interference it is strong, extraction accuracy is high.
Detailed description of the invention
Fig. 1 is flow diagram of the invention；
Fig. 2 is the digital picture obtained in the present invention；
Fig. 3 is in the present invention using the dark cross administrative division map obtained after Threshold segmentation；
Fig. 4 is that horizontal and vertical two-part signal is divided by origin of initial coordinate to dark cross region in the present invention
Figure, (a) are lateral part, (b) are longitudinal portion；
(the x that Fig. 5 is iteration 7 times in the embodiment of the present invention0,y0) and distance d value chart；
Fig. 6 is the curve synoptic diagram of the value of distance d iteration 7 times in the embodiment of the present invention；
Fig. 7 is dark cross searching coordinate (x obtained in the embodiment of the present invention0,y0) be (583.46,560.78) signal
Figure.
Specific embodiment
The present invention is further illustrated with reference to the accompanying drawings and examples.
A method of cross centre of figure being automatically extracted based on least square fitting iteration, steps are as follows:
(1) digital picture is obtained；Shown in following 2 figure of the collected electro-optic crystal conoscopic interference figure of the present embodiment, image district
Domain is 1200 × 1200 pixels；
(2) digital picture is subjected to Threshold segmentation, according to dark cross intensity profile, be arranged suitable gray value to image into
Row binaryzation extracts the dark cross region in digital picture, as shown in figure 3, color represents gray value in figure；
(3) initial coordinate (x in the dark cross region of iteration is set0,y0), that is, the geometric center position of digital picture is set
As initial coordinate (600,600)；
(4) with initial coordinate (600,600) for origin, by dark cross region segmentation at horizontal and vertical two parts, such as Fig. 4
It is shown, least square linear fit is weighted to horizontal and vertical two parts, obtains the two intersection point (x1,y1), weighting coefficient is
Amount related with each pixel gray value of image, if weighting coefficient is 1, linear fit is unrelated with gray value, otherwise Linear Quasi
Close it is related with gray value, the weighting coefficient that the present embodiment uses for the inverse of gray value square, to horizontal and vertical two parts into
Row weighted least-squares linear fit obtains the two intersection point (x1,y1) specific steps are as follows:
(41) y=a is calculated as to the fitting a straight line of lateral part1x+b1, then lateral least square fitting formula is as follows:
In formula, subscript i represents different pixels, xi,yiIt is the transverse and longitudinal of each pixel in image dark ten block lateral part domain respectively
Coordinate, ciIt is the corresponding weighting coefficient of each pixelIiFor the gray value of each pixel, m is the dark cross region of image
The total pixel number of lateral part；
(42) x=a is calculated as to the fitting a straight line of longitudinal portion2y+b2, then longitudinal least square fitting formula is as follows:
In formula, subscript i represents different pixels, xi,yiIt is the transverse and longitudinal of the dark each pixel in ten blocks longitudinal portion domain of image respectively
Coordinate, ciIt is the corresponding weighting coefficient of each pixelIiFor the gray value of each pixel, m is the dark cross region of image
The total pixel number of longitudinal portion；
(43) according to lateral part and longitudinal portion fitting a straight line equation, the two intersection point (x is acquired1,y1)；
A in formula1,b1It is the slope and intercept of lateral fitting a straight line, a respectively2,b2It is the slope of longitudinal fitting a straight line respectively
And intercept.
(5) (585.5651,567.3568) is found intersection at a distance from initial coordinate (600,600)And by (x1,y1) assignment is in initial coordinate (x0,y0),
I.e. by initial coordinate (x0,y0) it is revised as (585.5651,567.3568)；
(6) will distance d as iteration convergence condition, such as 0.01 pixel of d <, the iteration ends if meeting the condition of convergence are defeated
Initial coordinate (x out0,y0) centered on coordinate, otherwise go to step (4), the present embodiment iteration 7 times, this method convergence rate
Quickly, 0.01 pixel of distance d < after 5 iteration, as shown in figure 5 and figure 7；The dark cross searching coordinate (x that iteration obtains0,y0)
For (583.46,560.78), as shown in Figure 7.
The specific embodiment of the application above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
The limitation to the application protection scope therefore cannot be interpreted as.It should be pointed out that for those of ordinary skill in the art
For, under the premise of not departing from technical scheme design, various modifications and improvements can be made, these belong to this
The protection scope of application.
Claims (1)
1. a kind of method for automatically extracting cross centre of figure based on least square fitting iteration, which is characterized in that steps are as follows:
(1) digital picture is obtained；
(2) digital picture is subjected to Threshold segmentation, according to dark cross intensity profile, suitable gray value is set, two are carried out to image
Value extracts the dark cross region in digital picture；
(3) initial coordinate (x in the dark cross region of iteration is set0,y0), that is, the geometric center position conduct of digital picture is set
Initial coordinate (x0,y0)；
(4) with initial coordinate (x0,y0) it is origin, by dark cross region segmentation at horizontal and vertical two parts, to horizontal and vertical
Two parts are weighted least square linear fit, obtain the two intersection point (x1,y1)；
(5) (x is found intersection1,y1) and initial coordinate (x0,y0) distanceAnd by (x1,y1) assign
It is worth in initial coordinate (x0,y0)；
(6) will distance d as iteration convergence condition, the iteration ends if meeting the condition of convergence, export initial coordinate (x0,y0) make
For center coordinate, step (4) are otherwise gone to；
In the step (4), least square linear fit is weighted to horizontal and vertical two parts, obtains the two intersection point (x1,
y1) specific steps are as follows: (41) y=a is denoted as to the fitting a straight line of lateral part1x+b1, then lateral least square fitting formula is such as
Under:
In formula, subscript i represents different pixels, xi,yiIt is the transverse and longitudinal seat of each pixel in image dark ten block lateral part domain respectively
Mark, ciIt is the corresponding weighting coefficient of each pixelIiFor the gray value of each pixel, m is that the dark cross region of image is horizontal
To the total pixel number of part；
(42) x=a is denoted as to the fitting a straight line of longitudinal portion2y+b2, then longitudinal least square fitting formula is as follows:
In formula, subscript i represents different pixels, xi,yiIt is the transverse and longitudinal seat of the dark each pixel in ten blocks longitudinal portion domain of image respectively
Mark, ciIt is the corresponding weighting coefficient of each pixelIiFor the gray value of each pixel, m is that the dark cross region of image is vertical
To the total pixel number of part；
(43) according to lateral part and longitudinal portion fitting a straight line equation, the two intersection point (x is acquired1,y1)；
A in formula1,b1It is the slope and intercept of lateral fitting a straight line, a respectively2,b2Be respectively longitudinal fitting a straight line slope and cut
Away from.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610786591.9A CN106355613B (en) | 2016-08-31 | 2016-08-31 | The method for automatically extracting cross centre of figure based on least square fitting iteration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610786591.9A CN106355613B (en) | 2016-08-31 | 2016-08-31 | The method for automatically extracting cross centre of figure based on least square fitting iteration |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106355613A CN106355613A (en) | 2017-01-25 |
| CN106355613B true CN106355613B (en) | 2019-06-28 |
Family
ID=57857822
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610786591.9A Active CN106355613B (en) | 2016-08-31 | 2016-08-31 | The method for automatically extracting cross centre of figure based on least square fitting iteration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106355613B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107818583A (en) * | 2017-11-07 | 2018-03-20 | 歌尔股份有限公司 | Cross searching detection method and device |
| CN108090931B (en) * | 2017-12-13 | 2021-11-02 | 中国科学院光电技术研究所 | Anti-blocking and anti-interference marker identification and pose measurement method based on combination of circle and cross features |
| CN109754413A (en) * | 2018-12-27 | 2019-05-14 | 中国科学院长春光学精密机械与物理研究所 | A kind of Feisuo type dynamic interferometer bar graph method for registering |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN87207092U (en) * | 1987-11-18 | 1988-08-24 | 浙江大学 | Crystal opticle axle orientation device |
| CN104732553A (en) * | 2015-04-10 | 2015-06-24 | 大连理工大学 | Feature point extraction method based on multiple laser-assisted targets |
| CN104981105A (en) * | 2015-07-09 | 2015-10-14 | 广东工业大学 | Detecting and error-correcting method capable of rapidly and accurately obtaining element center and deflection angle |
| CN105066910A (en) * | 2015-08-21 | 2015-11-18 | 中国工程物理研究院激光聚变研究中心 | Electro-optic crystal Z axis deviation angle measurement device and measurement method |
-
2016
- 2016-08-31 CN CN201610786591.9A patent/CN106355613B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN87207092U (en) * | 1987-11-18 | 1988-08-24 | 浙江大学 | Crystal opticle axle orientation device |
| CN104732553A (en) * | 2015-04-10 | 2015-06-24 | 大连理工大学 | Feature point extraction method based on multiple laser-assisted targets |
| CN104981105A (en) * | 2015-07-09 | 2015-10-14 | 广东工业大学 | Detecting and error-correcting method capable of rapidly and accurately obtaining element center and deflection angle |
| CN105066910A (en) * | 2015-08-21 | 2015-11-18 | 中国工程物理研究院激光聚变研究中心 | Electro-optic crystal Z axis deviation angle measurement device and measurement method |
Non-Patent Citations (3)
| Title |
|---|
| CCD 图像法测量晶体光轴方向;沈为民等;《半导体光电》;20060830(第2006年第04期);全文 |
| Robustness of reweighted Least Squares Kernel Based Regression;Michiel Debruyne等;《Journal of Multivariate Analysis》;20090922;第101卷(第2期);全文 |
| 铌酸锂晶体电光调制系统的数字化研究;高文龙;《中国优秀硕士学位论文全文数据库 信息科技辑》;20140315(第2014年第03期);第42-43页 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106355613A (en) | 2017-01-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106295655B (en) | A kind of transmission line part extraction method for unmanned plane inspection image | |
| CN106355613B (en) | The method for automatically extracting cross centre of figure based on least square fitting iteration | |
| CN103630496B (en) | Based on the traffic video visibility detecting method of road surface apparent brightness and least square method | |
| CN102999749B (en) | Based on the securing band violation event intelligent detecting method of Face datection | |
| CN103528534B (en) | A kind of electric power line ice-covering thickness detection method based on image monitoring | |
| CN101382497A (en) | Visibility detecting method based on monitoring video of traffic condition | |
| CN107240079A (en) | A kind of road surface crack detection method based on image procossing | |
| CN105261018A (en) | Visibility detection method based on optical model and dark primary color priori theory | |
| CN106022341A (en) | High resolution optical remote sensing image post-disaster water body information extracting method and system | |
| CN103679677A (en) | Dual-model image decision fusion tracking method based on mutual updating of models | |
| CN103364793A (en) | SPOT5 image-based automatic water body extraction method | |
| CN109801236A (en) | A kind of photon point cloud denoising method based on mixed Gauss model | |
| CN102663720A (en) | Image splicing method based on minimum mean square error criterion | |
| CN103700078A (en) | Defogging method for image with few backgrounds containing fog | |
| CN104021527A (en) | Rain and snow removal method in image | |
| CN104615990A (en) | Method for automatically recognizing macula based on Huairou full-disk single-color image | |
| CN105809673A (en) | SURF (Speeded-Up Robust Features) algorithm and maximal similarity region merging based video foreground segmentation method | |
| Shao et al. | A kind of pavement crack detection method based on digital image processing | |
| Ming-wei et al. | Visibility detection based on traffic camera imagery | |
| CN106250880B (en) | Visual detection method for gradient of road before vehicle uphill | |
| CN103927523A (en) | Fog level detection method based on longitudinal gray features | |
| Ping et al. | Application of a sea surface temperature front composite algorithm in the Bohai, Yellow, and East China Seas | |
| CN103632373A (en) | Floc detection method combining three-frame differential higher-order statistics (HOS) with OTSU algorithm | |
| CN104574295B (en) | Adaptive threshold Image denoising algorithm | |
| CN108985307A (en) | A kind of Clean water withdraw method and system based on remote sensing image |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |