CN107067371A - Large format Leather Image joining method - Google Patents
Large format Leather Image joining method Download PDFInfo
- Publication number
- CN107067371A CN107067371A CN201710335284.3A CN201710335284A CN107067371A CN 107067371 A CN107067371 A CN 107067371A CN 201710335284 A CN201710335284 A CN 201710335284A CN 107067371 A CN107067371 A CN 107067371A
- Authority
- CN
- China
- Prior art keywords
- image
- straight line
- point
- calculated
- leather
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformation in the plane of the image
- G06T3/40—Scaling the whole image or part thereof
- G06T3/4038—Scaling the whole image or part thereof for image mosaicing, i.e. plane images composed of plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30124—Fabrics; Textile; Paper
Abstract
Large format Leather Image joining method, comprises the following steps：1) two images set gridiron pattern in corner and repeat region respectively；2) Harris Corner Detections are carried out successively；3) X-comers of repeat region are fitted；4) selection corresponding angles point is matched；5) matrixing is carried out to image characteristic point, calculates corresponding relation；6) coordinate projection conversion is carried out, you can the image spliced.The present invention has certain advantage on treatment effect and time efficiency.
Description
Technical field
The present invention relates to a kind of large format Leather Image joining method.
Background technology
Image mosaic is exactly by a width or several adjacent images for having an intersection, by series of algorithms, seamless spelling
Synthesize the process of piece image., it is necessary to carry out feature extraction, feature point detection etc. to image before image mosaic.The present invention
In Leather Image, regard gridiron pattern as reference, tessellated feature extracted, as image mosaic.
The general flow that leather substance is made is discharge, cutting, so the realization automated for leather, is primarily to obtain
The profile of leather.And in real operation, different animals leather is not of uniform size, Leather Image is disposably obtained to differ there is also difficulty or ease
Situation.The skin of larger animal such as ox, it is impossible to by disposably shooting acquisition image, then need to carry out image mosaic, obtained
Whole image.And the fur of meiofauna such as rabbit, then can be by once shooting acquisition image.Therefore, the leather of large format
The splicing of image and the design of contour outline extracting system have realistic meaning with realizing.
, it is necessary to carry out the registration of characteristics of image during image mosaic.Image registration is exactly, according to series of algorithms, to incite somebody to action
The process that multiple image is matched.The process of registration is exactly first to extract the characteristic point of two images, carries out Feature Points Matching, looks for
To characteristic point pair, further according to coordinate transform, images match is carried out.Point, line, surface are all conventional characteristics of image, wherein, extract figure
Characteristic point as in is most commonly used in use.Registration is carried out using the feature in image, the speed of image registration can be improved
Degree.Herein by tessellated feature detection, for image mosaic, it is possible to increase the efficiency of image mosaic.
The content of the invention
To overcome the larger animal skin of existing leather splicing can not be by disposably shooting the problem of obtaining image, this
Invention provides a kind of large format Leather Image stitching algorithm for effectively improving planning efficiency.
The large format Leather Image joining method of the present invention, comprises the following steps：
1), two images set gridiron pattern in corner and repeat region respectively.6 gridiron patterns are pasted in each image, up and down
Each three.Respectively positioned at the corner of image and the repeat region of image.
2) Harris Corner Detections, are carried out successively；
3), the X-comers of repeat region are fitted；
4), selection corresponding angles point is matched；
5) matrixing, is carried out to image characteristic point, corresponding relation is calculated；
6) coordinate projection conversion, the image spliced, are carried out.
Further, the step (3) comprises the steps of：
(3.1) domain is surrounded in setting：According to X-comers x, y-coordinate maximum and minimum value, domain is surrounded in setting.Surrounding
Straight line AB, BC, CD, the DA surrounded is an encirclement domain.An empty stack is set, for depositing single point.Fitting a straight line number
Measure as 0；
(3.2) the gridiron pattern length of side is calculated：Calculate the distance between coordinate two-by-two in angle point.Minimum range a is gridiron pattern
The length of side.Distance range be set as [0, a)；
(3.3) setting error s：Because between each angle point, position may recognize there is error.Therefore setting error s is needed.
With one side, initial BC is standard, is used as current straight line；
(3.4) judge whether stack is empty, empty, then into the 3.6th step, non-NULL then carries out the 3.9th step；
(3.5) each point is calculated to the distance of current straight line；
(3.6) point quantity of the distance in setting range is calculated, if only one of which point, into step 3.7.If have two and
Above point, distance range maximum subtracts a, into step 3.8；
(3.7) stacking.Whether judge fitting a straight line quantity is 0, if so, add a into distance range maximum, if it is not,
Into the 3.9th step；
(3.8) two points are chosen, straight line is calculated.Fitting a straight line quantity adds 1.Current straight line is used as using the straight line.If intending
The straight line quantity closed out is 5, into the 3.10th step, if into step 3.5；
(3.9) using current straight line as standard, because the straight line between gridiron pattern is all parallel relation.Can be according to current straight
Point in line and stack, calculated straight line a little.Point in stack is popped.Into step 3.6；
(3.10) according to the straight line fitted, it can determine each tessellated angular coordinate., can according to 4 straight lines
Tessellated scope is determined, therefore, it is possible to determine tessellated corner.
Further, the step (4) comprises the steps of：
(4.1) four angle points in four of two images repeating part tessellated maximum enclosure domains, as correspond to angle point；
(4.2) angle point is marked：Two images mutually corresponding angle point is marked.
Further, for step 4) selection correspondence X-comers are when matching, 4 pairs of chessboards in image repeat region
Only look for pair of horns point to be matched and marked in lattice, each pair gridiron pattern, obtain calculating 4 pairs of points required for conversion.
Further, the step (5) comprises the steps of：
(5.1) by 4 pairs of corresponding points calculate obtaining transformation matrix, formula is as follows；
(5.2) coefficient matrix LU is decomposed, to solve equation group, calculates m11-m32This 8 parameters；
Further, step (6) image mosaic, according to 8 transformation parameters calculated in step (5), construction splicing
Projective transformation matrix, image mosaic will be completed using projective transformation matrix on the first width image projection transformation to the second width image.
Splice projective transformation formula as follows：
The present invention technical concept be：For significantly leather, it is impossible to by disposably shooting acquisition image, then need into
Row image mosaic, obtains complete image, and the present invention regard gridiron pattern as reference, the tessellated spy of extraction when obtaining Leather Image
Levy a little, and carry out Feature Points Matching, calculate image mosaic projective transformation matrix, so as to realize the spelling of large format Leather Image
Connect.
It is an advantage of the invention that：The matching used time is short, and efficiency of algorithm is high, and algorithm stability is preferable, it is to avoid whole features are clicked through
Row matching.
Brief description of the drawings
Fig. 1 is total flow chart of the present invention
Embodiment
The invention will be further described below in conjunction with the accompanying drawings.
Reference picture 1, large format Leather Image joining method, comprises the following steps：
1), two images are respectively in corner and repeat region patch gridiron pattern.6 gridiron patterns are pasted in each image, up and down respectively
Three.Respectively positioned at the corner of image and the repeat region of image.
2) Harris Corner Detections, are carried out successively；
3), the X-comers of repeat region are fitted.Process is as follows：
(3.1) domain is surrounded in setting：According to X-comers x, y-coordinate maximum and minimum value, domain is surrounded in setting.Surrounding
Straight line AB, BC, CD, the DA surrounded is an encirclement domain.An empty stack is set, for depositing single point.Fitting a straight line number
Measure as 0；
(3.2) the gridiron pattern length of side is calculated：Calculate the distance between coordinate two-by-two in angle point.Minimum range a is gridiron pattern
The length of side.Distance range be set as [0, a)；
(3.3) setting error s：Because between each angle point, position may recognize there is error.Therefore setting error s is needed.
With one side, initial BC is standard, is used as current straight line；
(3.4) judge whether stack is empty, empty, then into the 3.6th step, non-NULL then carries out the 3.9th step；
(3.5) each point is calculated to the distance of current straight line；
(3.6) point quantity of the distance in setting range is calculated, if only one of which point, into step 3.7.If have two and
Above point, distance range maximum subtracts a, into step 3.8；
(3.7) stacking.Whether judge fitting a straight line quantity is 0, if so, add a into distance range maximum, if it is not,
Into the 3.9th step；
(3.8) two points are chosen, straight line is calculated.Fitting a straight line quantity adds 1.Current straight line is used as using the straight line.If intending
The straight line quantity closed out is 5, into the 3.10th step, if into step 3.5；
(3.9) using current straight line as standard, because the straight line between gridiron pattern is all parallel relation.Can be according to current straight
Point in line and stack, calculated straight line a little.Point in stack is popped.Into step 3.6；
(3.10) according to the straight line fitted, it can determine each tessellated angular coordinate., can according to 4 straight lines
Tessellated scope is determined, therefore, it is possible to determine tessellated corner.
4), selection corresponding angles point is matched.Process is as follows：
(4.1) four angle points in four of two images repeating part tessellated maximum enclosure domains, as correspond to angle point；
(4.2) angle point is marked：Two images mutually corresponding angle point is marked.
5) matrixing, is carried out to image characteristic point, corresponding relation is calculated.Process is as follows：
(5.1) by 4 pairs of corresponding points calculate obtaining transformation matrix, formula is as follows；
(5.2) coefficient matrix LU is decomposed, to solve equation group, calculates m11-m32This 8 parameters；
6) coordinate projection conversion, the image spliced, are carried out.
(6.1) according to 8 transformation parameters calculated in step (5), construction splicing projective transformation matrix is become using projection
Image mosaic will be completed on the first width image projection transformation to the second width image by changing matrix.Splice projective transformation formula as follows：
。
Claims (3)
1. large format Leather Image joining method, comprises the following steps：
1), two images set gridiron pattern in corner and repeat region respectively；6 gridiron patterns, up and down each three are pasted in each image
It is individual；Respectively positioned at the corner of image and the repeat region of image；
2) Harris Corner Detections, are carried out successively；
3), the X-comers of repeat region are fitted；Process is as follows：
(3.1) domain is surrounded in setting：According to X-comers x, y-coordinate maximum and minimum value, domain is surrounded in setting；All around around
Straight line AB, BC, CD, DA be one encirclement domain；An empty stack is set, for depositing single point；Fitting a straight line quantity is
0；
(3.2) the gridiron pattern length of side is calculated：Calculate the distance between coordinate two-by-two in angle point；Minimum range a is the gridiron pattern length of side；
Distance range be set as [0, a)；
(3.3) setting error s：Because between each angle point, position may recognize there is error；Therefore setting error s is needed；With one
Side, initial BC is standard, is used as current straight line；
(3.4) judge whether stack is empty, empty, then into the 3.6th step, non-NULL then carries out the 3.9th step；
(3.5) each point is calculated to the distance of current straight line；
(3.6) point quantity of the distance in setting range is calculated, if only one of which point, into step 3.7；If have two and more than
Point, distance range maximum subtracts a, into step 3.9；
(3.7) stacking；Whether judge fitting a straight line quantity is 0, if so, add a into distance range maximum, if it is not, into
3.9th step；
(3.8) two points are chosen, straight line is calculated；Fitting a straight line quantity adds 1；Current straight line is used as using the straight line；If fitting
Straight line quantity be 5, into the 3.10th step, if into step 3.5；
(3.9) using current straight line as standard, because the straight line between gridiron pattern is all parallel relation；Can according to current straight line and
Point in stack, calculated straight line a little；Point in stack is popped；Into step 3.6；
(3.10) according to the straight line fitted, it can determine each tessellated angular coordinate；According to 4 straight lines, it can determine
Tessellated scope, therefore, it is possible to determine tessellated corner；
4), selection corresponding angles point is matched；Process is as follows：
(4.1) four angle points in four of two images repeating part tessellated maximum enclosure domains, as correspond to angle point；
(4.2) angle point is marked：Two images mutually corresponding angle point is marked；
5) matrixing, is carried out to image characteristic point, corresponding relation is calculated；Process is as follows：
(5.1) by 4 pairs of corresponding points calculate obtaining transformation matrix, formula is as follows；
(5.2) coefficient matrix LU is decomposed, to solve equation group, calculates this 8 parameters of m11-m32；
6) coordinate projection conversion, the image spliced, are carried out.
2. large format Leather Image joining method as claimed in claim 1, it is characterised in that：For step 4) selection correspondence chess
When disk lattice angle point is matched, 4 pairs of gridiron patterns in image repeat region only look for pair of horns point to be matched in each pair gridiron pattern
And mark, obtain calculating 4 pairs of points required for conversion.
3. large format Leather Image joining method as claimed in claim 1 or 2, it is characterised in that：For step 6) image spelling
Connect, according to step 5) in 8 transformation parameters calculating, construction splicing projective transformation matrix, using projective transformation matrix by the
Piece image projective transformation is to completing image mosaic on the second width image.Splice projective transformation formula as follows：
。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710335284.3A CN107067371B (en) | 2017-05-12 | 2017-05-12 | Large-breadth leather image splicing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710335284.3A CN107067371B (en) | 2017-05-12 | 2017-05-12 | Large-breadth leather image splicing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107067371A true CN107067371A (en) | 2017-08-18 |
| CN107067371B CN107067371B (en) | 2020-05-26 |
Family
ID=59596261
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710335284.3A Active CN107067371B (en) | 2017-05-12 | 2017-05-12 | Large-breadth leather image splicing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107067371B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107680035A (en) * | 2017-09-29 | 2018-02-09 | 广东中星微电子有限公司 | A kind of parameter calibration method and device, server and readable storage medium storing program for executing |
| CN109146778A (en) * | 2018-07-02 | 2019-01-04 | 广西电网有限责任公司北海供电局 | Image split-joint method for workpiece size measurement |
| CN110570354A (en) * | 2019-09-10 | 2019-12-13 | 上海黑塞智能科技有限公司 | Strip chessboard calibration plate-based close-range image splicing method |
| CN110830783A (en) * | 2019-11-28 | 2020-02-21 | 歌尔科技有限公司 | VR image processing method and device, VR glasses and readable storage medium |
| CN112676183A (en) * | 2019-10-01 | 2021-04-20 | 格瑞玛股份公司 | Device and method for grading the quality of animal skins |
| CN113375555A (en) * | 2018-07-02 | 2021-09-10 | 广西电网有限责任公司北海供电局 | Power line clamp measuring method and system based on mobile phone image |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101110122A (en) * | 2007-08-31 | 2008-01-23 | 北京工业大学 | Large cultural heritage picture pattern split-joint method based on characteristic |
| CN104091324A (en) * | 2014-06-16 | 2014-10-08 | 华南理工大学 | Quick checkerboard image feature matching algorithm based on connected domain segmentation |
| CN104517276A (en) * | 2013-09-28 | 2015-04-15 | 沈阳新松机器人自动化股份有限公司 | Checker corner detection method |
-
2017
- 2017-05-12 CN CN201710335284.3A patent/CN107067371B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101110122A (en) * | 2007-08-31 | 2008-01-23 | 北京工业大学 | Large cultural heritage picture pattern split-joint method based on characteristic |
| CN104517276A (en) * | 2013-09-28 | 2015-04-15 | 沈阳新松机器人自动化股份有限公司 | Checker corner detection method |
| CN104091324A (en) * | 2014-06-16 | 2014-10-08 | 华南理工大学 | Quick checkerboard image feature matching algorithm based on connected domain segmentation |
Non-Patent Citations (2)
| Title |
|---|
| 张江雯: "皮革轮廓智能跟踪及测量的技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 郑洪波 等: "纹理去除的皮革图像颜色分类方法", 《计算机科学》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107680035A (en) * | 2017-09-29 | 2018-02-09 | 广东中星微电子有限公司 | A kind of parameter calibration method and device, server and readable storage medium storing program for executing |
| CN107680035B (en) * | 2017-09-29 | 2021-08-17 | 广东中星微电子有限公司 | Parameter calibration method and device, server and readable storage medium |
| CN109146778A (en) * | 2018-07-02 | 2019-01-04 | 广西电网有限责任公司北海供电局 | Image split-joint method for workpiece size measurement |
| CN113375555A (en) * | 2018-07-02 | 2021-09-10 | 广西电网有限责任公司北海供电局 | Power line clamp measuring method and system based on mobile phone image |
| CN110570354A (en) * | 2019-09-10 | 2019-12-13 | 上海黑塞智能科技有限公司 | Strip chessboard calibration plate-based close-range image splicing method |
| CN110570354B (en) * | 2019-09-10 | 2023-02-28 | 上海黑塞智能科技有限公司 | Strip chessboard calibration plate-based close-range image splicing method |
| CN112676183A (en) * | 2019-10-01 | 2021-04-20 | 格瑞玛股份公司 | Device and method for grading the quality of animal skins |
| CN110830783A (en) * | 2019-11-28 | 2020-02-21 | 歌尔科技有限公司 | VR image processing method and device, VR glasses and readable storage medium |
| CN110830783B (en) * | 2019-11-28 | 2021-06-01 | 歌尔光学科技有限公司 | VR image processing method and device, VR glasses and readable storage medium |
| WO2021103267A1 (en) * | 2019-11-28 | 2021-06-03 | 歌尔股份有限公司 | Vr image processing method and device, vr glasses, and readable storage medium |
| US11812008B2 (en) | 2019-11-28 | 2023-11-07 | Goertek Inc. | VR image processing method and device, VR glasses, and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107067371B (en) | 2020-05-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107067371A (en) | Large format Leather Image joining method | |
| CN106650630B (en) | A kind of method for tracking target and electronic equipment | |
| Shen et al. | Exemplar-based human action pose correction and tagging | |
| CN105354581B (en) | The color image feature extracting method of Fusion of Color feature and convolutional neural networks | |
| CN104346608A (en) | Sparse depth map densing method and device | |
| CN107562805A (en) | It is a kind of to scheme to search the method and device of figure | |
| CN103761765B (en) | Three-dimensional object model texture mapping method based on mapped boundaries optimization | |
| CN108225216A (en) | Structured-light system scaling method and device, structured-light system and mobile equipment | |
| CN108665491A (en) | A kind of quick point cloud registration method based on local reference | |
| CN105160686B (en) | A kind of low latitude various visual angles Remote Sensing Images Matching Method based on improvement SIFT operators | |
| CN101488224B (en) | Characteristic point matching method based on relativity measurement | |
| CN102004911B (en) | Method for improving accuracy of face identification | |
| Wang et al. | Image splicing detection based on convolutional neural network with weight combination strategy | |
| CN106022227A (en) | Gesture identification method and apparatus | |
| CN106295512A (en) | Many correction line indoor vision data base construction method based on mark and indoor orientation method | |
| CN107180436A (en) | A kind of improved KAZE image matching algorithms | |
| CN107680688A (en) | A kind of pelvic cavity emulation Minimally Invasive Surgery vision guided navigation verification method based on 3D printing | |
| CN106355607A (en) | Wide-baseline color image template matching method | |
| CN109816706A (en) | A kind of smoothness constraint and triangulation network equal proportion subdivision picture are to dense matching method | |
| CN106097354A (en) | A kind of combining adaptive Gauss Face Detection and the hand images dividing method of region growing | |
| CN105956544A (en) | Remote sensing image road intersection extraction method based on structural index characteristic | |
| CN110378336A (en) | Semantic class mask method, device and the storage medium of target object in training sample | |
| CN107133260A (en) | The matching and recognition method and device of a kind of landmark image | |
| CN110349209A (en) | Vibrating spear localization method based on binocular vision | |
| CN106355555A (en) | Image stitching 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20170818 Assignee: Hangzhou Anfeng Jiyue Cultural Creativity Co.,Ltd. Assignor: JIANG University OF TECHNOLOGY Contract record no.: X2022330000901 Denomination of invention: Mosaic Method of Large Scale Leather Images Granted publication date: 20200526 License type: Common License Record date: 20221228 Application publication date: 20170818 Assignee: Zhejiang Xinran Shoes Group Co.,Ltd. Assignor: JIANG University OF TECHNOLOGY Contract record no.: X2022330000882 Denomination of invention: Mosaic Method of Large Scale Leather Images Granted publication date: 20200526 License type: Common License Record date: 20221227 Application publication date: 20170818 Assignee: Hangzhou Zuolian Advertising Co.,Ltd. Assignor: JIANG University OF TECHNOLOGY Contract record no.: X2022330000922 Denomination of invention: Mosaic Method of Large Scale Leather Images Granted publication date: 20200526 License type: Common License Record date: 20221229 |
|
| EE01 | Entry into force of recordation of patent licensing contract | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20170818 Assignee: Zhejiang Xiudian Technology Co.,Ltd. Assignor: JIANG University OF TECHNOLOGY Contract record no.: X2023330000139 Denomination of invention: Method of stitching large leather images Granted publication date: 20200526 License type: Common License Record date: 20230311 |
|
| EE01 | Entry into force of recordation of patent licensing contract | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20170818 Assignee: Taizhou Luqiao Hongshi Plastic Factory Assignor: JIANG University OF TECHNOLOGY Contract record no.: X2023980038340 Denomination of invention: Method for stitching large leather images Granted publication date: 20200526 License type: Common License Record date: 20230720 Application publication date: 20170818 Assignee: Taizhou Printing Co.,Ltd. Assignor: JIANG University OF TECHNOLOGY Contract record no.: X2023980038345 Denomination of invention: Method for stitching large leather images Granted publication date: 20200526 License type: Common License Record date: 20230720 |
|
| EE01 | Entry into force of recordation of patent licensing contract |