CN109146778A - Image split-joint method for workpiece size measurement - Google Patents
Image split-joint method for workpiece size measurement Download PDFInfo
- Publication number
- CN109146778A CN109146778A CN201810714504.8A CN201810714504A CN109146778A CN 109146778 A CN109146778 A CN 109146778A CN 201810714504 A CN201810714504 A CN 201810714504A CN 109146778 A CN109146778 A CN 109146778A
- Authority
- CN
- China
- Prior art keywords
- image
- point
- algorithm
- workpiece
- splicing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000005259 measurement Methods 0.000 title claims abstract description 16
- 239000011159 matrix material Substances 0.000 claims abstract description 23
- 238000010276 construction Methods 0.000 claims abstract description 14
- 238000001514 detection method Methods 0.000 claims abstract description 13
- 230000009466 transformation Effects 0.000 claims abstract description 8
- 238000005070 sampling Methods 0.000 claims abstract description 5
- 238000012545 processing Methods 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 7
- 230000003287 optical effect Effects 0.000 claims description 3
- 238000011084 recovery Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 2
- 238000006243 chemical reaction Methods 0.000 description 4
- 230000004069 differentiation Effects 0.000 description 4
- 238000013178 mathematical model Methods 0.000 description 4
- 230000007704 transition Effects 0.000 description 4
- 230000009467 reduction Effects 0.000 description 3
- 230000000452 restraining effect Effects 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 239000003550 marker Substances 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
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
-
- G06T5/70—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- 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
- G06T2207/10024—Color image
Abstract
The invention discloses a kind of image split-joint methods for workpiece size measurement, select gridiron pattern scaling board known to size as reference, find key point (characteristic point) using the X-comers detection based on harris algorithm；Show that original image realizes essence matching with for the transformation matrix of coordinates for being registrated image key points using random sampling consistency RANSAC algorithm；Using suture algorithm and in view of exposure is different, by gray scale, construction, three aspect comprehensive considerations of exposure, best splicing line is obtained.In short, the present invention is a kind of good stability, high reliablity, the accurate method of splicing line, using the present invention can the more effective higher splicing for completing image of precision, the measurement for the later period for workpiece in high precision image lays a good foundation.
Description
Technical field
The invention belongs to technical field of image processing more particularly to a kind of image mosaic sides for workpiece size measurement
Method.
Background technique
Image mosaic is the research field of an increased popularity, it has become photograph cartography, computer vision, image
Hot topic in processing and computer graphics study.Image mosaic problem to be solved is normally behaved as by homogeneous
One seamless high-definition image of space overlap image configuration of series, it is with resolution ratio more higher than single image and more
The big visual field.
The first step for completing high precision image splicing is the higher picture method for registering of accuracy of selection, can just be found out in this way
Key point (i.e. characteristic point) in picture.Two images are exactly to splice after completing registration in next step.Traditional pixel weights splicing
Method is although efficient and convenient, but splices domain and be easy to appear the problems such as being overlapped ghost, does not have smoothing processing at splicing domain edge after all.
The principle of optimal stitching line is to calculate a splicing line in the overlapping region of two images, makes splicing line two sides pixel point
Not Lai Zi same piece image, belong to Dynamic Programming processing mathematical model.Scholars propose many calculating optimal stitching lines
Algorithm simultaneously achieves good results.
Duplaquet after absorbing a variety of suture algorithm characteristics, was proposed while in view of image in 1998
The algorithm of the color differentiation suture different with construction:
E (x, y)=Edif(x, y)2-γEedge(x, y)
Eedge(x, y)=min (g1(x, y), g2(x, y))
I1 and I2 respectively indicates two images to be spliced in formula, and E (x, y) dif represents the color for splicing domain in two images
Coloured silk difference, Eedge (x, y) indicate the architectural difference for splicing domain in two images.Wherein color differentiation E (x, y) dif is spelled with image
The average of the difference of the coordinate points in domain is connect to calculate.G1 and g2 is the gradient value of image to be spliced respectively, it is assumed that the knot of image
Structure difference Eedge (x, y) is calculated with the band lesser gradient of stitching image geometric deformation degree, and γ is fade factor in formula.
Mills proposed a kind of in view of there is exposure to distinguish the splicing line searching algorithm of operator in 2009:
E (x, y)=Ecolor(x, y)+Egeometry(x, y)
Ecolor (x, y) and Egeometry (x, y) is that color differentiation is different with construction respectively in formula, and I1 and I2 are respectively
The color degree of corresponding pixel points, g1 and g2 are the gradient values at picture registration partial dot (x, y) to be spliced, and α and β are exposure regions
Molecular group.Although this optimal stitching line calculation method solves the different bring splicing line result different problems of exposure, but
The flatness of image in splicing domain is had ignored, the problem that the α and β in this pattern are hardly resulted in, while also being had computational efficiency relatively low.
Summary of the invention
The technical problem to be solved in the present invention is to provide stability, and good, high reliablity, splicing line are accurately used for workpiece ruler
The image split-joint method of very little measurement.
In order to solve the above technical problems, the invention adopts the following technical scheme:
For the image split-joint method of workpiece size measurement, selects gridiron pattern scaling board known to size as reference, make
Key point (characteristic point) is found with the X-comers detection based on harris algorithm；It is calculated using random sampling consistency RANSAC
Method show that original image realizes essence matching with for the transformation matrix of coordinates for being registrated image key points；Using suture algorithm and consider
Exposure is different, by gray scale, construction, three aspect comprehensive considerations of exposure, obtains best splicing line.
The above-mentioned image split-joint method for workpiece size measurement, comprising the following steps:
<1>chessboard panel is put on workpiece, establishes image feature points of workpiece；
<2>X-comers in detection reference picture and image two images to be spliced；
<3>plane recovery is carried out to reference picture and image to be spliced；
<4>Feature Points Matching is carried out to the angle point that two images detect；
<5>mismatch operation is removed to matched characteristic point, improves matching accuracy；
<6>transformation matrix is calculated, two images are spliced；
<7>image co-registration, the processing of stitching image jointing line.
In harris algorithm, according to formula R (i, j)=det (M)-K* (trace (M))2To judge that this point is angle
Point, wherein det (M) is the value for seeking the determinant of M square matrix, and trace (M) is to seek the sum of element on M square matrix diagonal line, and K is exactly
One empirical value of harris Corner Detection Algorithm.
Plane recovery is to obtain tessellated all angle points, and known by screening to the characteristic point detected
Physical size between practical X-comers is established the list in image between X-comers and practical X-comers and is answered
Property, so that it may gridiron pattern is reverted to the state of vertical camera primary optical axis.
Suture utilizes following suture searching algorithm E (x, y)=α Egray(x, y)2+bEgeometry(x, y), E in formulagray
(x, y)2And Egeometry(x, y) respectively indicates the gray scale difference and construction difference of picture same section to be spliced, and a, b are gray area
Weighted values not different with construction；The process of splicing line searching algorithm is difference image, and it is poor to obtain picture registration part to be spliced
Score value extends splice point, splice point is linked up composition optimal stitching line.
, processing illumination interference not high for image mosaic precision for traditional images stitching algorithm and tradition splicing occur
Ghost problems, inventor establish it is a kind of for workpiece size measurement image split-joint method, select size known to chessboard
Case marker fixed board finds key point (characteristic point) as reference, using the X-comers detection based on harris algorithm；Using with
Machine sampling consistency RANSAC algorithm show that original image realizes essence matching with for the transformation matrix of coordinates for being registrated image key points；Make
With suture algorithm and in view of exposure is different, by gray scale, construction, three aspect comprehensive considerations of exposure, best splicing line is obtained.
Since the present invention is to calculate the actual size of workpiece using workpiece picture, therefore in calculating process, exist by scheming
Piece Pixel Dimensions therefore will have in workpiece picture scaling board known to a physical length to the conversion process of actual size
The transformational relation of Pixel Dimensions and actual size is acquired, so selecting gridiron pattern scaling board known to size as reference.In addition,
Since simple workpiece picture lacks in individuality a little and workpiece surface is coarse, present invention selection is made with the angle point on gridiron pattern scaling board
The characteristic point for converting and splicing for image.The position for accurately finding angle point inside gridiron pattern is of the invention primarily to ask
Topic.For this purpose, finding key point (characteristic point) using the X-comers detection based on harris algorithm.Carry out of characteristic point
Match, still is able to accurately extract key point in complex situations, there is higher stability and reliability.Nevertheless, by
It cannot be guaranteed the matched accuracy of key point in noise equal error, matching result screened, the present invention is taken out using random
Sample consistency RANSAC algorithm, obtain original image with for the transformation matrix of coordinates of image key points is registrated, to realize smart matching.
Splicing part uses optimal stitching line algorithm, searches for suture, makes grayscale information and construction poor by the coefficient factor of variation
It is different to be embodied to the greatest extent, and in view of exposure is different, gray scale, construction, three aspect comprehensive considerations of exposure make
Splicing line is more accurate.
In short, the present invention is a kind of good stability, high reliablity, the accurate method of splicing line, can more have using the present invention
The higher splicing for completing image of precision is imitated, the measurement for the later period for workpiece in high precision image is laid a good foundation.
Detailed description of the invention
Fig. 1 is the flow chart of image mosaic of the present invention.
Fig. 2 is the X-comers obtained after the present invention screens.
Fig. 3 is to apply the schematic diagram of the invention restored to gridiron pattern plane, and in figure: a left side is inclination gridiron pattern, and the right side is
Face gridiron pattern.
Fig. 4 is to apply the effect picture of the invention spliced, in figure: left is before splicing, and right is after splicing.
Specific embodiment
Lack due to being measured to workpiece size, on workpiece for splicing matched characteristic point, selects known to size
Gridiron pattern scaling board as reference, chessboard panel is put on workpiece, gridiron pattern scaling board and workpiece one to be measured are started auction
According to establishing image feature points of workpiece in case subsequent processing.
In order to complete precisely to splice, so that the workpiece size of measurement reaches the precision of submillimeter level, it is desirable that can accurately examine
The angle point on gridiron pattern scaling board is measured, the high image matching algorithm of choice accuracy is first had to, finds out the key point of image.This hair
Bright selection is using harris Corner Detection Algorithm to the detection of image key points, that is, characteristic point, and the algorithm is simple and easy, in reality
It is widely used in work.The basic principle of the algorithm is one target pixel points of selection, and chooses one with the mesh
The wicket centered on pixel is marked, calculates the grey scale change after window moves in any direction, and expressed with analytical form.
It is hereby achieved that substantially accurate angle point.It is specific as follows:
The gradient Iy in the direction gradient Ix and y in the direction x is calculated first.Predefined filter is established using fspecial function
Wave operator, syntax format are as follows: wherein type specifies the type of operator to h=fspecial (type, para), and para is specified corresponding
Parameter.The present invention uses gaussian Gassian low-pass filter, generates this window function of a 7*7, the standard value of filter
Sigma=2.It is put for each and calculates auto-correlation square M, according to the formula
R (i, j)=det (M)-K* (trace (M))2 (8)
To judge that this point is angle point.Wherein det (M) is the value for seeking the determinant of M square matrix, and trace (M) is to seek M
The sum of element on square matrix diagonal line, K are exactly an empirical value of harris Corner Detection Algorithm, generally take 0.04, K in the present invention
=0.04.Threshold value size simultaneously carries out non-maximum restraining, and the size of non-maximum restraining window can be tested more to be carried out really several times
Fixed, it is 3*3 that the present invention, which chooses non-maximum restraining window size,.The position of angle point is determined using find (result==1) sentence
It sets, and then counts the number of angle point.Utilize plot (posc, posr, ' r* ') mark the angle point shown.Utilize disp
(c) the angular coordinate c of acquisition is displayed on the screen by function, and coordinate c is saved as in a txt document.
But it can be due to camera inclination etc., so that the workpiece of shooting tilts on the image during shooting workpiece photo
Deformation, for the photograph angular distortion that camera orientation problem generates when eliminating shooting workpiece, therefore Corner Detection comes out and wants later
Restored by the plane that perspective transform carries out image to picture.Restoration methods be by being screened to the characteristic point detected,
It obtains tessellated all angle points (as shown in figure 3, having 7*4 totally 28 angle points in gridiron pattern), and known practical gridiron pattern angle
Physical size between point, establishes the homography in image between X-comers and practical X-comers, so that it may will
Gridiron pattern reverts to the state of vertical camera primary optical axis.
The characteristic point of two images after being restored using plane carries out corners Matching work, but due to many mistakes such as noise
Difference will affect the accuracy of H-matrix calculated result without accurately matching, can not keep away it cannot be guaranteed that the matched accuracy of key point
The appearance mismatch operation exempted from, so to continue to screen to matching result.It is thus desirable to a kind of data that can differentiate mistake pairing
Processing Algorithm, and H-matrix is computed correctly out in obtained accurate paired data.The present invention selects random sampling consistency
(RANSAC) smart matching is completed, this is a kind of to carrying out repeating extraction comprising error information data collection and calculate most suitably used number
According to the mathematical processing methods of conversion, RANSAC calculates embodiment source figure after excluding exterior point and obtaining the interior point with high accuracy again
Picture and with registration images match key point transformation matrix of coordinates.Specific mathematics process are as follows:
There are N in source picture and picture subject to registration to the key point pair correctly matched, while there is also incorrect
Pairing, might as well set the whole set of keypoints scanned is P, includes correct matching and deviation matching pair in this set P.It is real
Showing the conversion of key point in two images, corresponding pixel points should be realized by including the transition matrix H of m=8 numerical value,
I.e. the pixel of source picture obtains registration image slices vegetarian refreshments coordinate multiplied by matrix H.Since the change that image occurs can be more complicated,
Calculating a transition matrix H at least needs n=4 to calculate match point.Randomly choose 4 pairs of match points every time in this way to calculate
Matrix H.Detailed calculating process is as follows:
1, start to select n group match point from all matching set P.
2, initial conversion matrix H is obtained by the n group match point randomly selected.
3, the positional relationship with transition matrix is calculated to match point for N-n remaining in P, if distance is greater than a certain face
Dividing value T then brings outer point set into, other remaining match points bring the set u of interior point into, obtains interior point sum C.
4, above-mentioned steps 123 compute repeatedly K times.
5, most primary (i.e. c the is maximum) calculated result of point interior in k set of computations u is defined as correct matched group q.
6, final transition matrix H is calculated using set q obtained in the previous step.
Two images are exactly to splice after completing registration in next step.Due to reference picture and image capturing time to be spliced and
The effect of shadow of shooting is different, so that having apparent jointing line after the splicing of two pictures.Traditional pixel weights joining method
Although efficient and convenient, splicing domain is easy to appear the problems such as being overlapped ghost, does not have smoothing processing at splicing domain edge after all.To disappear
Except the ghost problem that traditional images fusion generates, this algorithm is solved using the method for optimal stitching line.Splicing principle of the invention
It is to calculate a splicing line in the overlapping region of two images, makes splicing line two sides pixel respectively from same piece image,
This belongs to the mathematical model of Dynamic Programming processing.Pixel locating for the optimal stitching line to be searched should be that two width pictures most connect
Close position, this approach can be quantified with several factors, and one is that nearby picture element point chromatic difference is minimum for splicing line, another
A is that splicing line description object geometrical construction nearby most communicates, and can use this two o'clock and carries out image mosaic.What it is by variation is
The number factor gives suture searching strip in view of exposure is different to enable grayscale information and structural differences to embody to the greatest extent
The image come, by gray scale, construction, three aspect comprehensive considerations of exposure, the splicing line made is more accurate and reliable.According to the above original
Reason designs and utilizes following suture searching algorithm:
E (x, y)=α Egray(x, y)2+bEgeometry(x, y)
E in formulagray(x, y)2And Egeometry(x, y) respectively indicate picture same section to be spliced gray scale difference (because
The problem of processing is gray level image, in this way can be to avoid color is calculated on three channels, reduces computation complexity) and construction
Difference, a, b are that different weighted values is distinguished and constructed to gray scale.If the exposure of image to be spliced differs greatly, fade factor
Value can be very big, the accuracy for doing exposure reduction at this time can be declined, and it is more quasi- to also result in two images gradient information in this way
Really, so the factor b value of gradient might as well be increased.To reduce error, first can handled apparent image is exposed
Reduce exposed image to a certain extent at simple reduction；Select mathematical model faster can be complete to reduce time cost
At reduction, it wouldn't use and calculate more complicated RANSAC mathematical model.Gray scale difference and structure are completed after obtaining factor ratio K
Make the different calculating factors.If depth of exposure is identical, it can simply enable grayscale information and structure differentiation coefficient identical, mathematics
On be expressed as the a=b if K=1, ab value in above formula are as follows:
K is image exposure coefficient value to be spliced in formula, and a when brightness ratio is 1, b song value is 0.5.
Above formula g1x and g2x are gradient value of each point in level in picture registration part to be spliced, g1y and g2y be to
Each coordinate points are then vertical direction gradients in stitching image intersection.
When calculating optimal stitching line, the mode of search is that each row is calculated from top to bottom.It is obtained in the first row
To after the starting point of suture, three points that a point is nearest in the distance of next line are respectively compared, which point difference minimum is then
As next extension point, splice point to the last a line of next line is then calculated with same procedure, is just completed at this time whole
The search process of a suture.Finally the splicing domain of image to be spliced can be divided into two pieces using optimal stitching line, two pieces of picture
Vegetarian refreshments respectively belongs to a picture.Very maximum probability moving picture can be retained in splicing line side in this way, avoid and spliced
The ghost problems that journey occurs.Certain optimal stitching line is also not the optimal algorithm for being fully solved problems, if certain point
Selection is inaccurate, and due to the algorithm extended downwardly, inevitably brings error to extension point and continues to final splicing
Point.If sufficiently high by the key point accuracy that image registration obtains, that is to say, that key point is the same point in practice,
So namely the smallest point of difference a, if splicing line is made of more key points it may be considered that this splicing line can
Sufficiently high by property, on the contrary then accuracy is insufficient, it is possible to pass through the pass near the key point or splicing line of splicing line approach
Measurement of the key point number as splicing line reliability.Registration and the matching of RANSAC key point secondary fine are calculated above by Euclidean
Method, it is ensured that the accuracy of key point.One so, which is defined, to the splicing line searching algorithm value near key point is greater than 0
Splicing line is set to be included in or nearby have most key points less than 1 coefficient (being set as 0.5).And in the mistake that splicing line extends
3 consecutive points might as well be become 5 consecutive points in journey with bigger possible close to key point, so the splicing that this research finally uses
Line search algorithmic procedure is difference image, obtains picture registration Partial Differential value to be spliced, extends splice point, splice point is connected
To form optimal stitching line.It is specific as follows:
Difference image: obtaining the difference value of picture registration part to be spliced, (preferably crucial from each point of the first row
Point) starting point as all possible splicing lines, the point for calculating its difference value, and choosing minimum difference value is defined as searching for seam
The starting point of zygonema；
Extend: extending since the aforementioned starting point found to next line, considers the adjacent 5 nearest points of next line, such as
Fruit certain point is that key point can compare this five difference values multiplied by 0.5 factor coefficient, selects the point of minimum value as under
The extension point of a line repeats the method until obtaining the splice point of last line pixel；
Optimal stitching line: the splice point determined in every a line, which is linked up, can be obtained by optimal stitching line.
Claims (5)
1. a kind of image split-joint method for workpiece size measurement, it is characterised in that gridiron pattern scaling board known to selection size
As reference, key point is found using the X-comers detection based on harris algorithm；Use random sampling consistency
RANSAC algorithm show that original image realizes essence matching with for the transformation matrix of coordinates for being registrated image key points；Use suture algorithm
And in view of exposure is different, by gray scale, construction, three aspect comprehensive considerations of exposure, best splicing line is obtained.
2. the image split-joint method according to claim 1 for workpiece size measurement, it is characterised in that including following step
It is rapid:
<1>chessboard panel is put on workpiece, establishes image feature points of workpiece；
<2>X-comers in detection reference picture and image two images to be spliced；
<3>plane recovery is carried out to reference picture and image to be spliced；
<4>Feature Points Matching is carried out to the angle point that two images detect；
<5>mismatch operation is removed to matched characteristic point, improves matching accuracy；
<6>transformation matrix is calculated, two images are spliced；
<7>image co-registration, the processing of stitching image jointing line.
3. the image split-joint method according to claim 1 for workpiece size measurement, it is characterised in that the harris
In algorithm, according to formula R (i, j)=det (M)-K* (trace (M))2To judge that this point is angle point, wherein det (M)
It is the value for seeking the determinant of M square matrix, trace (M) is to seek the sum of element on M square matrix diagonal line, and K is exactly that harris Corner Detection is calculated
One empirical value of method.
4. the image split-joint method according to claim 1 for workpiece size measurement, it is characterised in that: the plane is extensive
It is to obtain tessellated all angle points, and known practical X-comers by screening to the characteristic point detected again
Between physical size, establish the homography in image between X-comers and practical X-comers, so that it may by chess
Disk lattice revert to the state of vertical camera primary optical axis.
5. the image split-joint method according to claim 1 for workpiece size measurement, it is characterised in that: the suture
Utilize following suture searching algorithm E (x, y)=α Egray(x, y)2+bEgeometry(x, y), E in formulagray(x, y)2And Egeometry
(x, y) respectively indicates the gray scale difference and construction difference of picture same section to be spliced, and a, b are that gray scale is distinguished and constructed different
Weighted value；The process of splicing line searching algorithm is difference image, obtains picture registration Partial Differential value to be spliced, extends splicing
Splice point is linked up composition optimal stitching line by point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810714504.8A CN109146778A (en) | 2018-07-02 | 2018-07-02 | Image split-joint method for workpiece size measurement |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810714504.8A CN109146778A (en) | 2018-07-02 | 2018-07-02 | Image split-joint method for workpiece size measurement |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109146778A true CN109146778A (en) | 2019-01-04 |
Family
ID=64799664
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810714504.8A Pending CN109146778A (en) | 2018-07-02 | 2018-07-02 | Image split-joint method for workpiece size measurement |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109146778A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110738599A (en) * | 2019-10-14 | 2020-01-31 | 北京百度网讯科技有限公司 | Image splicing method and device, electronic equipment and storage medium |
| CN112037130A (en) * | 2020-08-27 | 2020-12-04 | 江苏提米智能科技有限公司 | Adaptive image splicing and fusing method and device, electronic equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130208997A1 (en) * | 2010-11-02 | 2013-08-15 | Zte Corporation | Method and Apparatus for Combining Panoramic Image |
| CN103438801A (en) * | 2013-09-16 | 2013-12-11 | 苏州天准精密技术有限公司 | Mobile platform type image measuring equipment and method |
| CN106530226A (en) * | 2016-10-25 | 2017-03-22 | 中源智人科技（深圳）股份有限公司 | Realization method for obtaining high-resolution high-definition industrial image |
| CN107067371A (en) * | 2017-05-12 | 2017-08-18 | 浙江工业大学 | Large format Leather Image joining method |
| US20180061006A1 (en) * | 2016-08-26 | 2018-03-01 | Multimedia Image Solution Limited | Method for ensuring perfect stitching of a subject's images in a real-site image stitching operation |
-
2018
- 2018-07-02 CN CN201810714504.8A patent/CN109146778A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130208997A1 (en) * | 2010-11-02 | 2013-08-15 | Zte Corporation | Method and Apparatus for Combining Panoramic Image |
| CN103438801A (en) * | 2013-09-16 | 2013-12-11 | 苏州天准精密技术有限公司 | Mobile platform type image measuring equipment and method |
| US20180061006A1 (en) * | 2016-08-26 | 2018-03-01 | Multimedia Image Solution Limited | Method for ensuring perfect stitching of a subject's images in a real-site image stitching operation |
| CN106530226A (en) * | 2016-10-25 | 2017-03-22 | 中源智人科技（深圳）股份有限公司 | Realization method for obtaining high-resolution high-definition industrial image |
| CN107067371A (en) * | 2017-05-12 | 2017-08-18 | 浙江工业大学 | Large format Leather Image joining method |
Non-Patent Citations (2)
| Title |
|---|
| 袁剑华: "广域监视中的运动目标检测相关技术研究与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 赵书睿: "全景图像拼接关键技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110738599A (en) * | 2019-10-14 | 2020-01-31 | 北京百度网讯科技有限公司 | Image splicing method and device, electronic equipment and storage medium |
| CN112037130A (en) * | 2020-08-27 | 2020-12-04 | 江苏提米智能科技有限公司 | Adaptive image splicing and fusing method and device, electronic equipment and storage medium |
| CN112037130B (en) * | 2020-08-27 | 2024-03-26 | 江苏提米智能科技有限公司 | Self-adaptive image stitching fusion method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104376548B (en) | A kind of quick joining method of image based on modified SURF algorithm | |
| CN103679636B (en) | Based on point, the fast image splicing method of line double characteristic | |
| CN100494886C (en) | Three-dimensional scanning system circular index point detection method | |
| CN103345755B (en) | A kind of Chessboard angular point sub-pixel extraction based on Harris operator | |
| CN108827316A (en) | Mobile robot visual orientation method based on improved Apriltag label | |
| US6671399B1 (en) | Fast epipolar line adjustment of stereo pairs | |
| CN104134200B (en) | Mobile scene image splicing method based on improved weighted fusion | |
| CN104599258B (en) | A kind of image split-joint method based on anisotropic character descriptor | |
| CN106485275A (en) | A kind of cover-plate glass of realizing positions, with liquid crystal display screen, the method fitted | |
| CN101394573B (en) | Panoramagram generation method and system based on characteristic matching | |
| CN105957007A (en) | Image stitching method based on characteristic point plane similarity | |
| CN105719318A (en) | Educational toy set and HSV based color identification method for Rubik's cube | |
| CN105046701B (en) | A kind of multiple dimensioned well-marked target detection method based on patterned lines | |
| CN105389774A (en) | Method and device for aligning images | |
| CN103902953B (en) | A kind of screen detecting system and method | |
| CN106663322A (en) | Identifying features | |
| CN103813095B (en) | Test chart and its application method | |
| CN109215016A (en) | A kind of recognition positioning method of coding maker | |
| CN102982334A (en) | Sparse parallax obtaining method based on target edge features and gray scale similarity | |
| CN109146778A (en) | Image split-joint method for workpiece size measurement | |
| JP5771423B2 (en) | Image color correction apparatus and image color correction method | |
| JP2015171143A (en) | Camera calibration method and apparatus using color-coded structure, and computer readable storage medium | |
| CN110060199A (en) | A kind of quick joining method of plant image based on colour and depth information | |
| JP5249813B2 (en) | MTF measurement chart, MTF measurement apparatus, and MTF measurement program | |
| CN104596486B (en) | Pose measuring method based on the rotationally symmetrical feature of target |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190104 |