CN106204544B - It is a kind of to automatically extract the method and system of mark point position and profile in image - Google Patents
It is a kind of to automatically extract the method and system of mark point position and profile in image Download PDFInfo
- Publication number
- CN106204544B CN106204544B CN201610499109.3A CN201610499109A CN106204544B CN 106204544 B CN106204544 B CN 106204544B CN 201610499109 A CN201610499109 A CN 201610499109A CN 106204544 B CN106204544 B CN 106204544B
- Authority
- CN
- China
- Prior art keywords
- point
- boundary
- image
- oval
- accurate
- 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
Classifications
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20032—Median filtering
Abstract
The method and system of mark point position and profile in image is automatically extracted the present invention relates to a kind of.The described method includes: obtaining original image;Median filter process is carried out to original image;Elliptical processing is fitted to the index point after filtering processing, obtains tentatively extracting result;The picpointed coordinate of the preliminary oval boundary point for extracting result is converted into polar coordinates;Several points are uniformly chosen on oval boundary and carry out accurate extraction process, finally obtain the precise information of index point.The present invention can more accurately automatically extract out the index point region in image by accurate extraction process on the basis of preliminary extraction process, and fit the parametric equation in the index point region;Extracted marker point profile can achieve sub-pixel, greatly improve the accuracy and stability of existing method.
Description
Technical field
The present invention relates to field of image processings, and in particular to one kind automatically extracts Circle in Digital Images shape index point position and profile
Method and system.
Background technique
A kind of important means that three-dimensional reconstruction is close shot three-dimensional reconstruction is carried out using stereoscopic vision.According to left and right two known to
The picpointed coordinate of correspondence same place in the image of internal and external orientation and the internal and external orientation of image utilize forward intersection
The three-dimensional coordinate changed the time can be obtained in principle.For convenience of the searching of corresponding same place, it is typically employed in object exterior pasting vision
The method of index point, this method is easy to operate, and calculation amount is few, and makees compared to the corresponding point of characteristic point is found on object
It is more accurate reliable for the method for same place.Due to imaging of the circular index point on the image of different perspectives be all it is oval or
Positive round is easy to be fitted, therefore visual beacon point mostly uses circle.It accurately extracts the boundary of index point on the image and center is to seek
Index point is looked for correspond to the basis of same place.
The method that index point is extracted in present image cannot mostly completely remove such as laser rays or other specks
Interference, the marker point profile line extracted is mostly pixel scale, and accuracy and correctness need to be improved.
Summary of the invention
The technical problem to be solved by the present invention is in view of the deficiencies of the prior art, provide one kind to automatically extract Circle in Digital Images
The method and system of shape index point position and profile.
The technical scheme to solve the above technical problems is that
A method of Circle in Digital Images shape index point position and profile are automatically extracted, is included the following steps:
Step 1, original image is obtained;
Step 2, median filter process is carried out to original image;
Step 3, elliptical processing is fitted to the index point after filtering processing, obtains tentatively extracting result;
Step 4, the picpointed coordinate of the preliminary oval boundary point for extracting result is converted into polar coordinates;
Step 5, several points are uniformly chosen on oval boundary and carry out accurate extraction process, finally obtain the accurate of index point
Data.
The beneficial effects of the present invention are: this method is by accurate extraction process, it can be on the basis of preliminary extraction process
The index point region in image is more accurately automatically extracted out, and fits the parametric equation in the index point region;It is extracted
Marker point profile can achieve sub-pixel, greatly improve the accuracy and stability of existing method.
Based on the above technical solution, the present invention can also be improved as follows.
Further, the step 2 is implemented as, and is chosen the image window that size is n*n, is carried out at median filtering
Reason, the gray value of the central pixel point of described image window is the intermediate value of gray value in described image window after processing;Traverse band
There is the original image of index point, successively chooses the image window that size is n*n and carry out median filter process.
Beneficial effect using above-mentioned further scheme is: can remove part using the image after median filter process and mark
Noise spot outside will point region, such as bright laser rays etc., rising mark point region.
Further, the specific implementation of the step 3 are as follows:
Step 3.1, edge detection is carried out to median filtering treated image using canny operator, calculates gradient change
Change the borderline region more than preset threshold;
Step 3.2, the boundary point detected, the perimeter C and area S of the borderline region of calculation flag point are traversed;
Step 3.3, circularity ε is calculated according to the perimeter C and area S, picks out circularity ε in section [0.6,1.0]
In boundary point set P1 { (x1, y1), (x2, y2) ..., (xn, yn) }, be the index point circular boundary point;
Step 3.4, using least square fitting elliptic equation, oval central coordinate of circle C (x is obtainedc,yc), it is transverse a, ellipse
Circle short axle b and transverse and image x-axis to angle α, the initial fitting of complement mark null ellipse;
Step 3.5, step 3.2-3.4 is repeated for each approximate index point, obtains tentatively extracting result.
Beneficial effect using above-mentioned further scheme is: carrying out edge detection, according to the boundary point detected, calculates side
The perimeter C and area S in battery limit (BL) domain calculate circularity ε, the circular boundary of available index point according to the perimeter C and area S
Point recycles least square fitting elliptic equation, obtains the oval center of circle, coordinate transverse a, ellipse short shaft b and transverse
With image x-axis to angle α, can be obtained and preliminary extract result.
Further, the step 4 is implemented as, by the boundary point of preliminary each index point for extracting result
Set P1 (x1, y1), (x2, y2) ..., (xn, yn) } picpointed coordinate be converted to the ellipse central coordinate of circle C (xc,yc)
For origin, the transverse a is the pixel collection { (θ under the polar coordinate system of axis positive direction1,r1),(θ2,r2),…,
(θn,rn)}。
Beneficial effect using above-mentioned further scheme is: the polar coordinates of the oval boundary point of initial fitting is obtained, after being
The Accurate Curve-fitting in face is prepared.
Further, the specific implementation of the step 5 are as follows:
Step 5.1, an angle variables θ is presetr, the θrFor the integer that can be divided exactly by 360, by θ every θrDegree is in institute
It states and takes a point on elliptical boundary, take 360/ θ altogetherrA boundary point obtains boundary point set CP { (x1,y1),(x2,y2),…};
Step 5.2, to 360/ θrA boundary point, on the direction θ, respectively each point (θi,ri) choose and itself phase
6 adjacent pixels, wherein 3 pixels are located in the elliptical boundary, 3 pixels are located at the elliptical boundary
Outside;
Step 5.3, to 360/ θrA boundary point is utilized respectively each point (θi,ri) 6 pixels distance it is oval
Distance { the r in the center of circle1,r2,r3,r4,r5,r6And pixel gray value { g1,g2,g3,g4,g5,g6Carried out according to determining algorithm
Accurate Curve-fitting obtains the accurate coordinates value (θ of each pointi,ri'), to obtain the set P2 of accurate oval boundary point
{(θ1,r1'), (θ2,r2') ... }.
Step 5.4, according to set the P2 { (θ1,r1'), (θ2,r2') ... }, using least square fitting elliptic equation,
Obtain the accurate coordinates C ' (x in the oval center of circlec', yc'), transverse a ', ellipse short shaft b ' and transverse and image x-axis
To angle α ', obtain the precise information of index point;
Step 5.5, step 5.1-5.4 is repeated for preliminary each index point for extracting result, completes original image subscript
The accurate extraction of will point.
Beneficial effect using above-mentioned further scheme is: realizing Accurate Curve-fitting, obtains the accurate coordinates data of index point.
It is a kind of to automatically extract the system of mark point position and profile in image, which is characterized in that including obtaining module, intermediate value
Module, preliminary extraction module, coordinate transferring and accurate extraction module is filtered;
The acquisition module, for obtaining original image;
The median filter process module, for carrying out median filter process to original image;
The preliminary extraction module obtains preliminary for being fitted elliptical processing to the index point after filtering processing
Extract result;
The coordinate transferring is sat for the picpointed coordinate of the preliminary oval boundary point for extracting result to be converted to pole
Mark;
The accurate extraction module carries out accurate extraction process for uniformly choosing several points on oval boundary, finally
Obtain the precise information of index point.
The beneficial effects of the present invention are: each module of this system carries out the accurate extraction process of index point by cooperating,
The index point region in image can be more accurately automatically extracted out on the basis of preliminary extraction process, and fits the mark
The parametric equation in point region;Extracted marker point profile can achieve sub-pixel, greatly improve the essence of existing method
Exactness and stability.
Further, the median filter process module includes selection unit and filter processing unit;
The selection unit, for choosing the image window that size is n*n;
The filter processing unit, for carrying out median filter process, the central pixel point of described image window after processing
Gray value be gray value in described image window intermediate value.
Beneficial effect using above-mentioned further scheme is: can remove part using the image after median filter process and mark
Noise spot outside will point region, such as bright laser rays etc., rising mark point region.
Further, the preliminary extraction module includes edge detection unit, perimeter areal calculation unit, circularity calculating list
Member and initial fitting unit;
The edge detection unit, for using canny operator to median filtering treated image progress edge detection,
Calculate the borderline region that change of gradient is more than preset threshold;
The perimeter areal calculation unit, for traversing the boundary point detected, the week of the borderline region of calculation flag point
Long C and area S;
The circularity computing unit picks out circularity ε for calculating circularity ε according to the perimeter C and area S
The set P1 { (x1, y1), (x2, y2) ..., (xn, yn) } of boundary point in section [0.6,1.0] is the circle of index point
Boundary point;
The initial fitting unit obtains oval central coordinate of circle C (x for utilizing least square fitting elliptic equationc,
yc), transverse a, ellipse short shaft b and transverse and image x-axis to angle α, the initial fitting of complement mark null ellipse,
It obtains tentatively extracting result.
Beneficial effect using above-mentioned further scheme is: carrying out edge detection, according to the boundary point detected, calculates side
The perimeter C and area S in battery limit (BL) domain calculate circularity ε, the circle of available approximation index point according to the perimeter C and area S
Boundary point recycles least square fitting elliptic equation, obtains the oval center of circle, coordinate transverse a, ellipse short shaft b and ellipse
Long axis and image x-axis to angle α, can be obtained and preliminary extract result.
Further, the coordinate transferring includes converting unit, for marking each preliminary for extracting result
The boundary point of will point set P1 (x1, y1), (x2, y2) ..., (xn, yn) } picpointed coordinate be converted to the oval circle
Heart coordinate C (xc,yc) it is origin, the transverse a is the pixel collection { (θ under the polar coordinate system of axis positive direction1,
r1),(θ2,r2),…,(θn,rn)}。
Beneficial effect using above-mentioned further scheme is: the polar coordinates of the oval boundary point of initial fitting is obtained, after being
The Accurate Curve-fitting in face is prepared.
Further, the accurate extraction module include take boundary dot element, take adjacent dot cells, boundary point fitting unit and
Accurate Curve-fitting unit;
It is described to take boundary dot element, for presetting an angle variables θr, the θrFor the integer that can be divided exactly by 360, press
θ is every θrDegree takes a point on the elliptical boundary, takes 360/ θ altogetherrA boundary point obtains boundary point set CP { (x1,
y1),(x2,y2),…};
It is described to take adjacent dot cells, for 360/ θrA boundary point, on the direction θ, respectively each point (θi,
ri) 6 pixels adjacent with itself are chosen, wherein 3 pixels are located in the elliptical boundary, 3 pixels are located at
Outside the elliptical boundary;
The boundary point fitting unit, for 360/ θrA boundary point is utilized respectively each point (θi,ri) 6
Distance { the r in the oval center of circle of pixel distance1,r2,r3,r4,r5,r6And pixel gray value { g1,g2,g3,g4,g5,g6Basis
Determining algorithm carries out Accurate Curve-fitting, obtains the accurate coordinates value (θ of each pointi,ri'), to obtain accurate elliptical side
Set the P2 { (θ of boundary's point1,r1'), (θ2,r2') ... };
The Accurate Curve-fitting unit, for according to set the P2 { (θ1,r1'), (θ2,r2') ... }, least square is utilized
Fitted ellipse equation obtains the accurate coordinates C ' (x in the oval center of circlec', yc'), transverse a ', ellipse short shaft b ' and ellipse
Long axis and image x-axis to angle α ', obtain the precise information of index point, complete the accurate extraction of the index point of original image.
Beneficial effect using above-mentioned further scheme is: realizing Accurate Curve-fitting, obtains the accurate coordinates data of index point.
Detailed description of the invention
Fig. 1 is a kind of method flow diagram for automatically extracting Circle in Digital Images shape index point position and profile of the present invention;
Fig. 2 is to read in raw video figure;
Fig. 3 is median filtering effect picture;
Fig. 4 is edge detection effect picture;
Fig. 5 is tentatively to extract single index point effect display figure;
Fig. 6 is that essence extracts elliptical schematic illustration;
Fig. 7 is that accurate logos point effect of extracting shows figure;
Fig. 8 is accurately to extract single index point effect picture;
Fig. 9 is accurately to extract single index point details effect picture;
Figure 10 is a kind of system block diagram for automatically extracting Circle in Digital Images shape index point position and profile of the present invention;
Figure 11 is linear filtering gradient template schematic diagram.
Specific embodiment
The principle and features of the present invention will be described below with reference to the accompanying drawings, and the given examples are served only to explain the present invention, and
It is non-to be used to limit the scope of the invention.
As shown in Figure 1, for a kind of method flow diagram for automatically extracting Circle in Digital Images shape index point position and profile of the present invention,
Include the following steps:
Step 1, original image is obtained;
In object exterior pasting visual beacon point, the index point is circle;By set up the equipment such as camera to object into
Row shooting, obtains the original image for having index point, as shown in Figure 2.
Step 2, median filter process is carried out to original image;
It chooses the image window that template size is n*n and carries out median filter process, processing formula is shown in formula (1).Time
The original image with index point is gone through, gray value of each window center pixel after median filter process is in the window
The intermediate value of gray value.
I=± 1 g (x, y)=med { f (x+i, y+i) }, ± 2, ± 3 (1)
The value of the n can be 3,5,7,9, preferred value 7.
The noise spot outside the index point region of part can be removed using the image after median filter process, such as is become clear sharp
Light, rising mark point region, as shown in Figure 3.
Step 3, it is fitted elliptical processing, obtains tentatively extracting result;
Edge detection is carried out using canny operator on this basis, calculates the boundary that change of gradient is more than preset threshold
Region, as shown in Figure 4.
The gradient is solved using linear filter method, and the present invention is using Sobe l operator, and template schematic diagram is as schemed
11.Evolution obtains the gradient of the point again for X-axis and the quadratic sum of y-axis direction gradient, and traversal image calculates the gradient of each pixel,
Obtain a gradient map.The calculation formula of the gradient is formula (2).
The preset threshold refers to that two threshold values are arranged in canny when detecting, and first threshold value can be set to 75, obtain
The pixel big to gradient;For the edge of building closure, second threshold can be set to 25, in two not connected endpoint pixels
Place, which filters out pixel according to second threshold, is closed edge as far as possible.
Traverse the boundary point detected, the perimeter C and area S of the borderline region of calculation flag point.
According to the perimeter C and area S, calculate circularity ε using formula (3), pick out circularity ε section [0.6,
1.0] the set P1 { (x1, y1), (x2, y2) ..., (xn, yn) } of the boundary point in is the border circular areas of the index point.
According to the set P1 { (x1, y1), (x2, y2) ..., (xn, yn) } for the approximate circle boundary point picked out, utilize
Least square fitting elliptic equation obtains elliptical central coordinate of circle C (xc,yc), major semiaxis length a, semi-minor axis length b, ellipse length
Axis and image x-axis to angle α, the initial fitting of complement mark null ellipse, as shown in Figure 5.
Above procedure is repeated, the preliminary extraction result of the index point of original image is obtained.
Step 4, the picpointed coordinate of the preliminary oval boundary point for extracting result is converted into polar coordinates;
By it is described it is preliminary extract result each index point boundary point set P1 (x1, y1), (x2, y2) ...,
(xn, yn) } picpointed coordinate be converted to the ellipse central coordinate of circle C (xc,yc) be origin, the transverse a be axis just
Pixel collection { (θ under the polar coordinate system in direction1,r1),(θ2,r2),…,(θn,rn), meet formula (4), wherein e is ellipse
Round eccentricity is calculated according to formula (5).
The polar coordinates of the oval boundary point of the initial fitting of all index points on original image are obtained, are subsequent accurate quasi-
Conjunction is prepared.
Step 5, several points are uniformly chosen on oval boundary and carry out accurate extraction process, finally obtain the accurate of index point
Data;
Preset an angle variables θr, the θrFor the integer that can be divided exactly by 360, by θ every θrDegree is described elliptical
A point is taken on boundary, takes 360/ θ altogetherrA boundary point obtains boundary point set CP { (x1,y1),(x2,y2),…};
To 360/ θrA boundary point, on the direction θ, respectively each point (θi,ri) choose 6 adjacent with itself
Pixel, wherein 3 pixels are located in the elliptical boundary, 3 pixels are located at outside the elliptical boundary, such as Fig. 6
It is shown;
It is utilized respectively each point (θi,ri) the oval center of circle of 6 pixels distance distance { r1,r2,r3,r4,r5,r6And
Gray value { the g of pixel1,g2,g3,g4,g5,g6Fitting formula (6) cubic polynomial, matrix equation can be listed according to 6 points
(7), formula (8) are represented by, by the parameter of cubic polynomial can be calculated using formula (9) according to the principle of least square, to three
Order polynomial carries out secondary derivation, can obtain the polynomial inflection point formula (10), obtain (θi,ri'), (θi,riIt ') is the angle
The position of the more accurate oval boundary point of degree.
G=A1 × r3+A2×r2+A3×r+A4 (6)
G=R × A (8)
A=(RT×R)-1×RT×G (9)
As shown in figure 8, accurately to extract single index point effect picture;As shown in figure 9, thin accurately to extract single index point
Save effect picture.
To 360/ θrEach of a boundary point carries out above-mentioned calculating, to obtain the collection of accurate oval boundary point
Close P2 { (θ1,r1'), (θ2,r2') ... }.
According to set the P2 { (θ1,r1'), (θ2,r2') ... }, it using least square fitting elliptic equation, obtains described
The accurate coordinates C ' (x in the oval center of circlec', yc'), transverse a ', ellipse short shaft b ' and transverse and image x-axis to angle
α ' obtains the precise information of index point;
The θrValue can be 1,2,3,4,5,6, preferred value 3.
It is repeated the above process tentatively to extract each approximate index point of result, completes the essence of index point on original image
It really extracts, if Fig. 7 is accurate extraction logos point effect display figure.
It as shown in Figure 10, is a kind of system block diagram for automatically extracting Circle in Digital Images shape index point position and profile of the present invention,
Including obtaining module, median filter process module, preliminary extraction module, coordinate transferring and accurate extraction module;It is described to obtain
Modulus block, for obtaining original image;The median filter process module, for carrying out median filter process to original image;
The preliminary extraction module obtains tentatively extracting result for being fitted elliptical processing to the index point after filtering processing;
The coordinate transferring, for the picpointed coordinate of the preliminary oval boundary point for extracting result to be converted to polar coordinates;The essence
True extraction module carries out accurate extraction process for uniformly choosing several points on oval boundary, finally obtains the essence of index point
Exact figures evidence.
The median filter process module includes selection unit and filter processing unit;The selection unit, for choosing
Size is the image window of n*n;The filter processing unit, for carrying out median filter process, described image window after processing
Central pixel point gray value be gray value in described image window intermediate value.
The preliminary extraction module include edge detection unit, perimeter areal calculation unit, circularity computing unit and just
Walk fitting unit;The edge detection unit, for using canny operator to median filtering treated image progress edge inspection
It surveys, calculates the borderline region that change of gradient is more than preset threshold;The perimeter areal calculation unit is detected for traversing
Boundary point, the perimeter C and area S of the borderline region of calculation flag point;The circularity computing unit, for according to the perimeter
C and area S calculate circularity ε, pick out boundary point of the circularity ε in section [0.6,1.0] set P1 (x1, y1),
(x2, y2) ..., (xn, yn) }, it is the circular boundary point of index point;The initial fitting unit, for quasi- using least square
Elliptic equation is closed, oval central coordinate of circle C (x is obtainedc,yc), transverse a, ellipse short shaft b and transverse and image x-axis to
Angle α, the initial fitting of complement mark null ellipse obtain tentatively extracting result.
The coordinate transferring includes converting unit, for by it is described it is preliminary extract result each index point side
Boundary's point set P1 (x1, y1), (x2, y2) ..., (xn, yn) } picpointed coordinate be converted to the oval central coordinate of circle C
(xc,yc) it is origin, the transverse a is the pixel collection { (θ under the polar coordinate system of axis positive direction1,r1),(θ2,
r2),…,(θn,rn)}。
The accurate extraction module includes taking boundary dot element, taking adjacent dot cells, boundary point fitting unit and accurately intend
Close unit;It is described to take boundary dot element, preset an angle variables θr, the θrFor the integer that can be divided exactly by 360, by θ every
θrDegree takes a point on the elliptical boundary, takes 360/ θ altogetherrA boundary point obtains boundary point set CP { (x1,y1),(x2,
y2),…};It is described to take adjacent dot cells, for 360/ θrA boundary point, on the direction θ, respectively each point (θi,
ri) 6 pixels adjacent with itself are chosen, wherein 3 pixels are located in the elliptical boundary, 3 pixels are located at
Outside the elliptical boundary;The boundary point fitting unit, for 360/ θrA boundary point is utilized respectively each point
(θi,ri) the oval center of circle of 6 pixels distance distance { r1,r2,r3,r4,r5,r6And pixel gray value { g1,g2,g3,
g4,g5,g6According to determining algorithm progress Accurate Curve-fitting, obtain the accurate coordinates value (θ of each pointi,ri'), to obtain
Set the P2 { (θ of accurate ellipse boundary point1,r1'), (θ2,r2') ... };The Accurate Curve-fitting unit, for according to the collection
Close P2 { (θ1,r1'), (θ2,r2') ... }, using least square fitting elliptic equation, the accurate coordinates in the oval center of circle are obtained
C ' (xc', yc'), transverse a ', ellipse short shaft b ' and transverse and image x-axis to angle α ', obtain the essence of index point
Exact figures evidence completes the accurate extraction of the index point of original image.
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all in spirit of the invention and
Within principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
Claims (6)
1. a kind of automatically extract the method for mark point position and profile in image, which comprises the steps of:
Step 1, original image is obtained;
Step 2, median filter process is carried out to original image;
Step 3, elliptical processing is fitted to the index point after filtering processing, obtains tentatively extracting result;
Step 4, the picpointed coordinate of the preliminary oval boundary point for extracting result is converted into polar coordinates;
Step 5, several points are uniformly chosen on oval boundary and carry out accurate extraction process, finally obtain the perfect number of index point
According to;
The specific implementation of the step 3 are as follows:
Step 3.1, edge detection is carried out to median filtering treated image using canny operator, it is super calculates change of gradient
Cross the borderline region of preset threshold;
Step 3.2, the boundary point detected, the perimeter C and area S of the borderline region of calculation flag point are traversed;
Step 3.3, circularity ε is calculated according to the perimeter C and area S, picks out circularity ε in section [0.6,1.0]
The set P1 { (x1, y1), (x2, y2) ..., (xn, yn) } of boundary point is the circular boundary point of the index point;
Step 3.4, using least square fitting elliptic equation, oval central coordinate of circle C (x is obtainedc,yc), it is transverse a, oval short
Axis b and transverse and image x-axis to angle α, the initial fitting of complement mark null ellipse;
Step 3.5, step 3.2-3.4 is repeated for each index point, obtains tentatively extracting result;
The specific implementation of the step 5 are as follows:
Step 5.1, an angle variables θ is presetr, the θrFor the integer that can be divided exactly by 360, by θ every θrDegree is described ellipse
A point is taken on round boundary, takes 360/ θ altogetherrA boundary point obtains boundary point set CP { (x1,y1),(x2,y2),…};
Step 5.2, to 360/ θrA boundary point, on the direction θ, respectively each point (θi,ri) choose it is adjacent with itself
6 pixels, wherein 3 pixels are located in the elliptical boundary, 3 pixels are located at outside the elliptical boundary;
Step 5.3, to 360/ θrA boundary point is utilized respectively each point (θi,ri) the oval center of circle of 6 pixels distance
Distance { r1,r2,r3,r4,r5,r6And pixel gray value { g1,g2,g3,g4,g5,g6Carried out accurately according to determining algorithm
Fitting, obtains the accurate coordinates value (θ of each pointi,ri'), to obtain set the P2 { (θ of accurate oval boundary point1,
r1'), (θ2,r2') ... };
Step 5.4, according to set the P2 { (θ1,r1'), (θ2,r2') ... }, it using least square fitting elliptic equation, obtains
The accurate coordinates C ' (x in the ellipse center of circlec', yc'), transverse a ', ellipse short shaft b ' and transverse and image x-axis to
Angle α ' obtains the precise information of index point;
Step 5.5, step 5.1-5.4 is repeated for preliminary each index point for extracting result, completes index point on original image
Accurate extraction.
2. a kind of according to claim 1 automatically extract the method for mark point position and profile in image, which is characterized in that institute
Being implemented as step 2 is stated, the image window that size is n*n is chosen, carries out median filter process, described image window after processing
The gray value of the central pixel point of mouth is the intermediate value of gray value in described image window;Traversal has the original image of index point,
It successively chooses the image window that size is n*n and carries out median filter process.
3. a kind of according to claim 1 automatically extract the method for mark point position and profile in image, which is characterized in that institute
State being implemented as step 4, by it is described it is preliminary extract result each index point boundary point set P1 (x1, y1),
(x2, y2) ..., (xn, yn) picpointed coordinate be converted to the ellipse central coordinate of circle C (xc,yc) it is origin, it is described oval long
Axis a is the pixel collection { (θ under the polar coordinate system of axis positive direction1,r1),(θ2,r2),…,(θn,rn)}。
4. a kind of automatically extract the system of mark point position and profile in image, which is characterized in that including obtaining module, intermediate value filter
Wave processing module, preliminary extraction module, coordinate transferring and accurate extraction module;
The acquisition module, for obtaining original image;
The median filter process module, for carrying out median filter process to original image;
The preliminary extraction module is tentatively extracted for being fitted elliptical processing to the index point after filtering processing
As a result;
The coordinate transferring, for the picpointed coordinate of the preliminary oval boundary point for extracting result to be converted to polar coordinates;
The accurate extraction module carries out accurate extraction process for uniformly choosing several points on oval boundary, finally obtains
The precise information of index point;
The preliminary extraction module includes edge detection unit, week into areal calculation unit, circularity computing unit and preliminary quasi-
Close unit;
The edge detection unit, for carrying out edge detection, calculating to median filtering treated image using canny operator
Change of gradient is more than the borderline region of preset threshold out;
The perimeter areal calculation unit, for traversing the boundary point detected, the perimeter C of the borderline region of calculation flag point and
Area S;
The circularity computing unit picks out circularity ε in area for calculating circularity ε according to the perimeter C and area S
Between boundary point in [0.6,1.0] set P1 { (x1, y1), (x2, y2) ..., (xn, yn) }, be the circle of the index point
Boundary point;
The initial fitting unit obtains oval central coordinate of circle C (x for utilizing least square fitting elliptic equationc,yc), it is ellipse
Circle long axis a, ellipse short shaft b and transverse and image x-axis to angle α, the initial fitting of complement mark null ellipse obtains just
Step extracts result;
The accurate extraction module includes taking boundary dot element, taking adjacent dot cells, boundary point fitting unit and Accurate Curve-fitting list
Member;
It is described to take boundary dot element, for presetting an angle variables θr, the θrIt is every by θ for the integer that can be divided exactly by 360
Every θrDegree takes a point on the elliptical boundary, takes 360/ θ altogetherrA boundary point obtains boundary point set CP { (x1,y1),
(x2,y2),…};
It is described to take adjacent dot cells, for 360/ θrA boundary point, on the direction θ, respectively each point (θi,ri) choosing
6 pixels adjacent with itself are taken, wherein 3 pixels are located in the elliptical boundary, 3 pixels are located at described ellipse
Outside round boundary;
The boundary point fitting unit, for 360/ θrA boundary point is utilized respectively each point (θi,ri) 6 pixels
Distance { the r in the oval center of circle of point distance1,r2,r3,r4,r5,r6And pixel gray value { g1,g2,g3,g4,g5,g6According to determination
Algorithm carry out Accurate Curve-fitting, obtain the accurate coordinates value (θ of each pointi,ri'), to obtain accurate oval boundary point
Set P2 { (θ1,r1'), (θ2,r2') ... };
The Accurate Curve-fitting unit, for according to set the P2 { (θ1,r1'), (θ2,r2') ... }, least square fitting is utilized
Elliptic equation obtains the accurate coordinates C ' (x in the oval center of circlec', yc'), transverse a ', ellipse short shaft b ' and transverse
With image x-axis to angle α ', obtain the precise information of index point.
5. a kind of according to claim 4 automatically extract the system of mark point position and profile in image, which is characterized in that institute
Stating median filter process module includes selection unit and filter processing unit;
The selection unit, for choosing the image window that size is n*n;
The filter processing unit, for carrying out median filter process, the ash of the central pixel point of described image window after processing
Angle value is the intermediate value of gray value in described image window.
6. a kind of according to claim 4 automatically extract the system of mark point position and profile in image, which is characterized in that institute
Stating coordinate transferring includes converting unit, for by it is described it is preliminary extract result index point boundary point set P1
(x1, y1), (x2, y2) ..., (xn, yn) } picpointed coordinate be converted to the ellipse central coordinate of circle C (xc,yc) it is origin,
The transverse a is the pixel collection { (θ under the polar coordinate system of axis positive direction1,r1),(θ2,r2),…,(θn,rn)}。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610499109.3A CN106204544B (en) | 2016-06-29 | 2016-06-29 | It is a kind of to automatically extract the method and system of mark point position and profile in image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610499109.3A CN106204544B (en) | 2016-06-29 | 2016-06-29 | It is a kind of to automatically extract the method and system of mark point position and profile in image |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106204544A CN106204544A (en) | 2016-12-07 |
CN106204544B true CN106204544B (en) | 2019-04-05 |
Family
ID=57463452
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610499109.3A Active CN106204544B (en) | 2016-06-29 | 2016-06-29 | It is a kind of to automatically extract the method and system of mark point position and profile in image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106204544B (en) |
Families Citing this family (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106650717B (en) * | 2016-12-17 | 2020-07-28 | 复旦大学 | Accurate positioning method for round object with thickness interference |
CN107369140B (en) * | 2017-06-29 | 2020-05-19 | 大连理工大学 | Method for extracting center of high-precision target ball in unstructured environment |
CN107578051B (en) * | 2017-09-14 | 2020-09-29 | 西华大学 | Detection and identification method of annular coding mark points |
CN108205395B (en) * | 2018-01-16 | 2021-03-23 | 安徽慧视金瞳科技有限公司 | Method for accurately positioning center coordinates of calibration points |
CN108921836A (en) | 2018-06-28 | 2018-11-30 | 京东方科技集团股份有限公司 | A kind of method and device for extracting eye fundus image mark |
CN109238268A (en) * | 2018-08-01 | 2019-01-18 | 北京理工大学 | The optimal external ellipses recognition method of irregular small feature loss navigation centroid |
CN108827316B (en) * | 2018-08-20 | 2021-12-28 | 南京理工大学 | Mobile robot visual positioning method based on improved Apriltag |
CN109410268B (en) * | 2018-11-06 | 2020-06-23 | 温州高视雷蒙光电科技有限公司 | Method and system for determining circle center of concentric ring of corneal topography |
CN111354047B (en) * | 2018-12-20 | 2023-11-07 | 精锐视觉智能科技(上海)有限公司 | Computer vision-based camera module positioning method and system |
CN110009655B (en) * | 2019-02-12 | 2020-12-08 | 中国人民解放军陆军工程大学 | Eight-direction three-dimensional operator generation and use method for stereo image contour enhancement |
CN110057296A (en) * | 2019-04-11 | 2019-07-26 | 河海大学 | Monitoring method is measured based on the indoor slope model test local displacement that frame per second is extracted |
CN110345877B (en) * | 2019-07-03 | 2020-10-27 | 西安交通大学 | Method for measuring aperture and pitch of tube plate |
CN110458809B (en) * | 2019-07-16 | 2020-11-17 | 西安工程大学 | Yarn evenness detection method based on sub-pixel edge detection |
CN110378417B (en) * | 2019-07-19 | 2023-10-31 | 东南大学 | Method for acquiring construction boundary |
CN110490846B (en) * | 2019-07-30 | 2023-01-06 | 武汉睿芯特种光纤有限责任公司 | Method and device for testing geometric dimension of optical fiber image |
CN111079772B (en) * | 2019-12-18 | 2024-01-26 | 深圳科瑞技术股份有限公司 | Image edge extraction processing method, device and storage medium |
CN112116581B (en) * | 2020-09-23 | 2023-09-08 | 中国科学院物理研究所 | Method and device for acquiring atomic position in atomic imaging |
CN112698380A (en) * | 2020-12-16 | 2021-04-23 | 南京大学 | Beam section processing method suitable for low-energy proton beam under strong background noise |
CN112258660B (en) * | 2020-12-23 | 2021-03-30 | 恒信东方文化股份有限公司 | Processing method and system of virtual dressing image |
CN114087989B (en) * | 2021-11-19 | 2023-09-22 | 江苏理工学院 | Method and system for measuring three-dimensional coordinates of circle center of positioning hole of automobile cylinder workpiece |
CN117115242B (en) * | 2023-10-17 | 2024-01-23 | 湖南视比特机器人有限公司 | Identification method of mark point, computer storage medium and terminal equipment |
CN117351526B (en) * | 2023-12-05 | 2024-03-22 | 深圳纯和医药有限公司 | Intravascular ultrasound image automatic identification method for intima |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101311963A (en) * | 2008-06-17 | 2008-11-26 | 东南大学 | Round mark point center picture projection point position acquiring method for positioning video camera |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
SE0101932D0 (en) * | 2001-05-31 | 2001-05-31 | Astrazeneca Ab | Pharmaceutical combinations |
-
2016
- 2016-06-29 CN CN201610499109.3A patent/CN106204544B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101311963A (en) * | 2008-06-17 | 2008-11-26 | 东南大学 | Round mark point center picture projection point position acquiring method for positioning video camera |
Non-Patent Citations (4)
Title |
---|
"三维相机检校场标志点图像坐标提取方法";刘凤英等;《测绘工程》;20150731;第24卷(第7期);论文第48-52页 |
"基于双目立体视觉的GEO非合作目标位姿测量方法研究";薛强;《中国优秀硕士学位论文全文数据库 信息科技辑》;20140415;论文第37页 |
"视频测量影像序列椭圆形人工目标点快速识别和跟踪方法";刘祥磊等;《测绘学报》;20150630;第44卷(第6期);论文第664-666页 |
"面向惯性约束聚变实验靶图像的快速椭圆检测";吴倩等;《中国图象图形学报》;20140131;第19卷(第1期);论文第77-83页 |
Also Published As
Publication number | Publication date |
---|---|
CN106204544A (en) | 2016-12-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106204544B (en) | It is a kind of to automatically extract the method and system of mark point position and profile in image | |
CN109785379B (en) | Method and system for measuring size and weight of symmetrical object | |
CN107833181B (en) | Three-dimensional panoramic image generation method based on zoom stereo vision | |
CN109000557B (en) | A kind of nuclear fuel rod pose automatic identifying method | |
CN108416791A (en) | A kind of monitoring of parallel institution moving platform pose and tracking based on binocular vision | |
CN106780590A (en) | The acquisition methods and system of a kind of depth map | |
CN108171787A (en) | A kind of three-dimensional rebuilding method based on the detection of ORB features | |
CN110956661B (en) | Method for calculating dynamic pose of visible light and infrared camera based on bidirectional homography matrix | |
CN107767456A (en) | A kind of object dimensional method for reconstructing based on RGB D cameras | |
CN104933434A (en) | Image matching method combining length between perpendiculars (LBP) feature extraction method and surf feature extraction method | |
CN108447022B (en) | Moving target joining method based on single fixing camera image sequence | |
CN106996748A (en) | A kind of wheel footpath measuring method based on binocular vision | |
CN106228513A (en) | A kind of Computerized image processing system | |
CN105913013A (en) | Binocular vision face recognition algorithm | |
CN112669280B (en) | Unmanned aerial vehicle inclination aerial photography right-angle image control point target detection method based on LSD algorithm | |
CN107677274A (en) | Unmanned plane independent landing navigation information real-time resolving method based on binocular vision | |
CN103993548A (en) | Multi-camera stereoscopic shooting based pavement damage crack detection system and method | |
CN106570899A (en) | Target object detection method and device | |
CN108154536A (en) | The camera calibration method of two dimensional surface iteration | |
CN113689401A (en) | Method and device for detecting diameter of crystal bar of czochralski silicon single crystal furnace | |
CN107358628A (en) | Linear array images processing method based on target | |
CN105606123A (en) | Method for automatic correction of digital ground elevation model for low-altitude aerial photogrammetry | |
CN109859137A (en) | A kind of irregular distortion universe bearing calibration of wide angle camera | |
CN106534833A (en) | Space and time axis joint double-viewpoint three dimensional video stabilizing method | |
CN103700082B (en) | Image split-joint method based on dual quaterion relative orientation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right |
Effective date of registration: 20200810 Address after: 430000 room 15, floor 5, building 6, Guanggu power energy conservation and environmental protection industrial park, No. 308, Guanggu Avenue, Donghu New Technology Development Zone, Wuhan City, Hubei Province (intersection of Fenghuangyuan Middle Road) Patentee after: WUHAN ZHONGGUAN AUTOMATION TECHNOLOGY Co.,Ltd. Address before: 210012, room 7, floor 1, 706 West Spring Road, Yuhuatai District, Jiangsu, Nanjing, China Patentee before: NANJING ZHONGGUAN SOFTWARE TECHNOLOGY Co.,Ltd. |
|
TR01 | Transfer of patent right |