CN102944191B - Method and device for three-dimensional vision measurement data registration based on planar circle target - Google Patents
Method and device for three-dimensional vision measurement data registration based on planar circle target Download PDFInfo
- Publication number
- CN102944191B CN102944191B CN201210494953.9A CN201210494953A CN102944191B CN 102944191 B CN102944191 B CN 102944191B CN 201210494953 A CN201210494953 A CN 201210494953A CN 102944191 B CN102944191 B CN 102944191B
- Authority
- CN
- China
- Prior art keywords
- circle
- target
- splicing
- dimensional
- plane
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 75
- 238000005259 measurement Methods 0.000 title claims abstract description 65
- 239000011159 matrix material Substances 0.000 claims abstract description 47
- 239000013598 vector Substances 0.000 claims description 21
- 238000004364 calculation method Methods 0.000 claims description 15
- 238000005457 optimization Methods 0.000 claims description 12
- 238000010586 diagram Methods 0.000 description 7
- 238000013519 translation Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 239000000523 sample Substances 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
Landscapes
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses a method for three-dimensional vision measurement data registration based on a planar circle target. The method includes fixing and placing an object to be tested to a reasonable position capable of being observed by a marked binocular vision system, and building a global coordinate system; placing the planar circle target before the object to be tested, moving the marked binocular vision system, shooting elliptic images formed by measurement positions before and after the movement; extracting ellipses from the shot planar circle target images, fitting elliptic equations, and rebuilding circle characteristics of the circle target under the two local measurement coordinates; and constructing and optimizing objective functions according to the circle characteristics and solving a registration matrix. The invention further discloses a device for the three-dimensional vision measurement data registration based on the planar circle target. According to the method and the device, the problem of failure of three-dimensional data registration under the condition that a target portion is shielded in the prior art can be solved, the accuracy of three-dimensional measurement data registration is guaranteed, and reliability of the three-dimensional measurement data registration is improved.
Description
Technical Field
The invention relates to a three-dimensional vision measurement splicing technology, in particular to a three-dimensional vision measurement data splicing method and device based on a plane circle target.
Background
In order to measure the three-dimensional shape of a large-size measured object, the surface of the large-size measured object is generally divided into a plurality of sub-areas, each sub-area is measured from a plurality of visual angles, and each local measurement data is spliced to a global coordinate system. The accuracy of three-dimensional data splicing determines the accuracy which can be achieved by the visual measurement of the three-dimensional appearance of a large-size measured object, so that the method has important practical significance for the research of a three-dimensional data splicing method.
At present, there are three main methods for three-dimensional data splicing:
first, the measurement range is extended by a large-scale device such as a precision platform, a theodolite, a laser tracker, etc., and the measurement device used in this method is expensive and has a limited measurement range;
secondly, marking points are pasted in a public view field of two adjacent measurements of the measuring device, and a splicing matrix is obtained by utilizing three non-collinear points, so that the method has the defects that the work of pasting and removing the marks is more complicated, and the surface of a measured object is damaged;
the third one is an Iterative Closest Point (ICP) algorithm, but this method has the problems of large Iterative computation amount, long operation time, and the like, and is not suitable for the measured object with surface curvature variation not rich.
The method based on the plane target overcomes the defects of the three methods, and the method utilizes the characteristic points provided by the plane target in the adjacent twice measurement public view field to calculate the splicing matrix, so that the method is convenient to operate, high in precision and wide in application prospect. At present, a planar target based on grid characteristics is commonly used, and a splicing matrix is obtained by matching grid angular points, so that three-dimensional data splicing is realized. However, in practical applications, when a planar target portion is shielded, corner points of the squares are often mismatched, so that the three-dimensional data splicing fails.
Disclosure of Invention
In view of the above, the main objective of the present invention is to provide a method and an apparatus for splicing three-dimensional vision measurement data based on a planar circular target, which solve the problem of failure in splicing three-dimensional data when the planar circular target is partially shielded in the prior art, ensure the precision of splicing three-dimensional measurement data, and improve the reliability of splicing three-dimensional measurement data.
In order to achieve the purpose, the technical scheme of the invention is realized as follows:
the invention provides a three-dimensional vision measurement data splicing method based on a plane circle target, which comprises the following steps:
fixing and placing the measured object at a reasonable position which can be observed by a calibrated binocular vision system, and establishing a global coordinate system;
placing the planar circular target in front of the measured object, moving the calibrated binocular vision system, and shooting an elliptical image formed by the measuring positions of the planar circular target before and after movement;
extracting an ellipse from the shot plane circle target image, fitting an ellipse equation, and reconstructing circle characteristics of the plane circle target under the two-time local measurement coordinate system;
and constructing an optimized objective function according to the circle characteristics, and solving a splicing matrix.
In the above scheme, the circle feature includes a center coordinate and a normal vector of the circle plane.
In the above scheme, the fitting ellipse equation is: an ellipse equation is fitted with a direct least squares method combined with a Random Sample Consensus algorithm (RANSAC, Random Sample Consensus).
In the above scheme, the circle feature of the reconstructed planar circle target in the previous and subsequent local measurement coordinate systems is as follows:
and reconstructing the circle feature under the local measurement coordinate system by utilizing a geometrically intersected circle feature three-dimensional reconstruction method based on an elliptic equation, a camera parameter matrix and the radius value of the circle feature on the plane circle target which are fitted by combining a direct least square method of RANSAC.
In the above scheme, the constructing an optimized objective function according to the circle feature to solve the splicing matrix includes:
constructing an optimization objective function by utilizing the identity of circle characteristics of the reconstructed front and rear measurement positions in space, and obtaining the ratio of inner points and outer points when fitting an elliptic equation according to RANSAC; and setting the weight corresponding to the matched edge point set, and calculating the high-precision splicing matrix by utilizing a Levenberg-Marquardt (Levenberg-Marquardt) nonlinear optimization method.
The invention also provides a device for realizing the three-dimensional vision measurement data splicing based on the plane circle target, which comprises an elliptical image acquisition module, a circle reconstruction characteristic module and a splicing matrix calculation module; wherein,
the elliptical image acquisition module is used for acquiring elliptical images formed by measuring positions of the planar circular target before and after movement;
the circle feature reconstruction module is used for extracting an ellipse from the shot plane circle target ellipse image, fitting an ellipse equation and reconstructing circle features of the plane circle target in a local measurement coordinate system of the front and back times;
and the splicing matrix calculation module is used for constructing an optimized objective function according to the circle characteristics and solving the splicing matrix.
In the scheme, the elliptical image acquisition module, the circle reconstruction characteristic module and the splicing matrix calculation module are arranged in a binocular vision system.
The three-dimensional data splicing method and device based on the planar circle target utilize the identity of circle features and rich edge information of circles; in the process of obtaining the splicing matrix, the accuracy of fitting the elliptic equation is improved by detecting the elliptic edge point set with high accuracy, and then the obtaining accuracy of the splicing matrix is improved. In addition, the invention can improve the accuracy and reliability of three-dimensional data splicing by utilizing the advantage that the circle feature can be accurately extracted and matched under the condition that the circle feature is partially shielded, thereby solving the problem that the three-dimensional data splicing fails under the condition that the circle feature is partially shielded by adopting a plane grid target method in the prior art, ensuring the accuracy of three-dimensional measurement data splicing and improving the reliability of three-dimensional measurement data splicing.
In addition, the circular target is not easily affected by partial shielding, so the circular target is more suitable for being applied to various complex field environments.
Drawings
FIG. 1 is a schematic diagram of a three-dimensional visual measurement data stitching model based on a planar circular target according to the present invention;
FIG. 2 is a schematic flow chart of a three-dimensional vision measurement data splicing method based on a planar circular target according to the present invention;
FIG. 3 is a schematic structural diagram of the planar circular target of the present invention;
FIG. 4 is a schematic diagram of the structure of the device for splicing the three-dimensional vision measurement data based on the planar circular target according to the present invention;
FIG. 5 is a diagram of an experimental system according to an embodiment of the present invention;
FIG. 6 is a three-dimensional cloud of points of the local measurement locations of the present invention;
FIG. 7 is a graph comparing the results of the present invention and the ICP method;
FIG. 8 is a schematic view of a scene where a right camera takes a partially occluded image;
FIG. 9 is a schematic diagram of a three-dimensional point set of a grid target under a partial occlusion condition;
FIG. 10 is a matching graph of three-dimensional point sets of grid targets at positions 1 and 2 under partial occlusion.
Detailed Description
For better understanding of the present invention, first, a basic principle of three-dimensional vision measurement data stitching is introduced, fig. 1 is a schematic diagram of a three-dimensional vision measurement data stitching model based on a planar circular target, as shown in fig. 1, a camera coordinate system O is respectively established at a local measurement position k and a position k +1ckxckyckzckAnd Ock+1xck+1yck+1zck+1The global coordinate system is established at the camera coordinate system O of position 1c1xc1yc1zc1The following steps of (1); during the camera movement of the two previous local measurements, the planar circular target 11 remains stationary, that is: the circle features on the planar circle target 11 are invariant under the global coordinate system; here, the circle feature includes a center coordinate and a normal vector of a circle plane;
assuming a circular characteristic ofIn the camera coordinate system OckxckyckzckAnd Ock+1xck+1yck+1zck+1Respectively have the results ofAndsuppose a camera coordinate system OckxckyckzckAnd Ock+1xck+1yck+1zck+1Has a mosaic matrix of Mk+1，k，Mk+1，kBy orthogonal 3 x 3 rotation matrix Rk+1，kAnd a 3 × 1 translation vector tk+1，kThe composition is expressed by formula (1):
in the process of measuring the position k to the position k +1 locally, according to the identity of the circle feature of the planar circle target in the global coordinate system, the following relationship exists:
pk+1＝Rk+1，k·pk+tk+1，k (3)
r can be calculated by the formula (2) and the formula (3)k+1，k、tk+1，kFurther, the transformation matrix M of the camera coordinate system can be calculated according to the formula (1)k+1，k。
Taking the coordinate system of the camera at the position 1 as a global measurement coordinate system, and transforming M the measurement data at the position k +12，1M3，2…Mk+1，kCan be unified to a global coordinate system Oc1xc1yc1zc1The following steps. Similarly, the local measurement data of the three-dimensional scanning probe at all positions can be unified to the global coordinate system Oc1xc1yc1zc1Then, three-dimensional shape data of the whole measured object is obtained, and the splicing of the three-dimensional vision measurement data is completed; the three-dimensional scanning measuring head comprises the two cameras and a projector and is used for acquiring three-dimensional shape data of a measured object. Typically, the binocular vision system 12 includes two cameras for capturing the stitched image.
According to the basic principle of three-dimensional vision measurement data splicing, the basic idea of the invention is as follows: fixing and placing the measured object at a reasonable position which can be observed by a calibrated binocular vision system, and establishing a global coordinate system; placing the planar circular target in front of the measured object, moving the calibrated binocular vision system, and shooting an elliptical image formed by the measuring positions of the planar circular target before and after movement; extracting an ellipse from the shot plane circle target image, fitting an ellipse equation, and reconstructing circle characteristics of the plane circle target under the two-time local measurement coordinate system; and constructing an optimized objective function according to the circle characteristics to solve a splicing matrix.
The technical solution of the present invention is further elaborated below with reference to the drawings and the specific embodiments.
Fig. 2 is a schematic flow chart of a three-dimensional vision measurement data stitching method based on a planar circular target according to the present invention, and as shown in fig. 2, the method includes:
step 201, fixing and placing a measured object at a reasonable position which can be observed by a calibrated binocular vision system, and establishing a global coordinate system;
specifically, the Calibration method used for calibrating the binocular vision system can be referred to "adaptive New Technique for Camera Calibration [ J ] (IEEE trans. pattern analysis and Machine Intelligence, nov.2000)" by z.y.zhang;
the reasonable positions where the measured object is placed in the calibrated binocular vision system to be observed are as follows: dividing the measured object into several subregions, placing the calibrated binocular vision system at the position 1, and setting the global coordinate system under the coordinate system of one camera in the binocular vision system and recording as Oc1xc1yc1zc1(ii) a Here, the binocular vision system comprises two cameras;
specifically, the number of sub-regions into which the object to be measured is divided is determined according to the shape of the object to be measured, and the purpose is to splice the overall shape of the object to be measured through a binocular vision system; preferably, there is one-third overlap between each two subregions.
Step 202, placing the planar circular target in front of the object to be measured, moving the calibrated binocular vision system, and shooting an elliptical image formed by the measuring positions of the planar circular target before and after movement;
here, the planar circular target is placed in front of the measured object subregion;
specifically, fig. 3 is a schematic diagram of a planar circular target used in the embodiment of the present invention, as shown in fig. 3, a plane of the planar circular target is a square, concentric circles in the target are composed of three circles, diameters of the circles are 95mm, 65mm and 35mm in sequence from large to small, and a side length of the square is 125 mm; the concentric circles are used for accurately extracting the coordinates of the center of the space circle and the normal vector of the circle plane; and the outer square is used for determining a target coordinate system and calculating the correspondence of the circle characteristic points through nonlinear optimization.
Here, because the wholeness of the circle feature on the plane circle target, the plane circle target has the advantage that the circle feature can be accurately extracted under the condition that the part is invisible, therefore, only the part of the circle target feature needs to be shot when the mobile binocular system shoots, and the convenience of the movement of the binocular vision system is greatly improved. Since the projected image of the circle in space is an ellipse, taking a planar circular target will result in an elliptical image.
Step 203, extracting an ellipse from the shot plane circle target image, fitting an ellipse equation, and reconstructing circle characteristics of the plane circle target in the previous and subsequent local measurement coordinate systems;
wherein the circle feature comprises a circle center coordinate and a normal vector of a circle plane;
specifically, the calculation in FIG. 1Andfor an image formed by a camera imaging plane of a plane circular target in a binocular vision system, processing the image by methods such as sub-pixel edge detection based on an improved canny operator to obtain an ellipse point set, filtering out disordered points by using an elliptical feature, and fitting an ellipse equation by combining a direct least square method of RANSAC; under the condition of partial shielding, the edge points of the shielding object extracted from the edge do not belong to the elliptical circular arc.
Here, a point on the elliptical arc is referred to as an inner point; the edge points of the obstruction are called outer points; because the accuracy of fitting the ellipse equation by the direct least square method is reduced due to the existence of the outer points, the inner points and the outer points of the edge ellipse are distinguished by using RANSAC, so that the influence of the outer points is removed, and the accuracy of fitting the ellipse is improved.
Specifically, the circle feature of the reconstructed planar circle target under the two previous and subsequent local measurement coordinate systems is as follows: based on an elliptic equation, a camera parameter matrix and a radius value of a circle feature on a plane circle target which are fitted by combining a direct least square method of RANSAC, a circle feature under a local measurement coordinate system is reconstructed by utilizing a geometrically-intersected circle feature three-dimensional reconstruction method, namely: finding the upper circle of the planar circle target under the local measurement coordinate system of the position kThe three-dimensional circle center coordinate and the normal vector of the feature are respectivelyIn the same way, obtain
204, constructing an optimized objective function according to the circle characteristics, and solving a splicing matrix;
here, an optimization objective function is constructed using the identity in space of the circular features of the preceding and following measurement positions;
specifically, the step is to calculate the rotation matrix R and the translation vector t shown in fig. 1; because the rotation matrix has three degrees of freedom, at least two pairs of vectors need to be selected, and the rotation matrix R is obtained by calculation by utilizing the two pairs of vectors and cross multiplication vectors thereof; however, the linear calculation method has low accuracy and is generally used as an initial value for nonlinear calculation and global optimization.
When more than three pairs of vectors are known, the rotation matrix R can be obtained by least squares linear regression, and the calculation formula is shown in formula (4):
R＝Nk+1·Nk T·(Nk·Nk T)-1 (4)
in the formula (4), the first and second groups, <math>
<mrow>
<msub>
<mi>N</mi>
<mi>k</mi>
</msub>
<mo>=</mo>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mover>
<mi>v</mi>
<mo>→</mo>
</mover>
<mrow>
<mi>k</mi>
<mo>,</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>,</mo>
<msub>
<mover>
<mi>v</mi>
<mo>→</mo>
</mover>
<mrow>
<mi>k</mi>
<mo>,</mo>
<mn>2</mn>
</mrow>
</msub>
<mo>,</mo>
<mo>.</mo>
<mo>.</mo>
<mo>.</mo>
<mo>.</mo>
<mo>,</mo>
<msub>
<mover>
<mi>v</mi>
<mo>→</mo>
</mover>
<mrow>
<mi>k</mi>
<mo>,</mo>
<mi>n</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msup>
<mo>;</mo>
</mrow>
</math> <math>
<mrow>
<msub>
<mi>N</mi>
<mrow>
<mi>k</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>=</mo>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mover>
<mi>v</mi>
<mo>→</mo>
</mover>
<mrow>
<mi>k</mi>
<mo>+</mo>
<mn>1</mn>
<mo>,</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>,</mo>
<msub>
<mover>
<mi>v</mi>
<mo>→</mo>
</mover>
<mrow>
<mi>k</mi>
<mo>+</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
</mrow>
</msub>
<mo>,</mo>
<mo>.</mo>
<mo>.</mo>
<mo>.</mo>
<mo>.</mo>
<mo>,</mo>
<msub>
<mover>
<mi>v</mi>
<mo>→</mo>
</mover>
<mrow>
<mi>k</mi>
<mo>+</mo>
<mn>1</mn>
<mo>,</mo>
<mi>n</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msup>
<mo>;</mo>
</mrow>
</math> n is the number of pairs of normal vectors of the matching spatial circular planes, and n > 3.
After the rotation matrix R is obtained by equation (4), the translation vector t can be obtained by equation (5):
t＝pk+1-R·pk (5)
based on a normal vector pair of the matched space circular plane, the rotation matrix R is expressed in a Rodrigues vector form, and the proportion value of an inner point and an outer point is obtained when an ellipse is calculated according to RANSAC fitting to set the vector pairAndcorresponding weight wiThe least square problem of constructing the rotation matrix R by optimization is shown in equation (6):
and (3) calculating an optimized rotation matrix R by using a Levenberg-Marquardt nonlinear optimization method for the formula (6).
On the basis of the translation vector t calculated by the formula (5) and the rotation matrix R calculated by the formula (6), an optimization objective function is constructed by using the matched edge point set of the space circle and the center coordinates of the space circle, and a splicing matrix is calculated, as shown in the formula (7):
similarly, the rotation matrix R is represented in the form of Rodrigues vectors and an ellipse is fitted according to RANSACSetting matched edge point set p according to the ratio of inner point to outer point obtained in equationkAnd pk+1Corresponding weight wiCalculating a high-precision rotation matrix R and a translation vector t by using a Levenberg-Marquardt nonlinear optimization method (M.Galassi, J.Davies, J.Theiler, G.Jungman, M.Booth, and F.Rossi.GNU Scientific Library Reference Manual, Network Theory, Aug.2006); and then, calculating an accurate splicing matrix according to the formula (1) to complete the splicing of the three-dimensional vision measurement data.
Fig. 4 is a schematic view of a composition structure of a device for implementing three-dimensional vision measurement data splicing based on a planar circular target according to the present invention, as shown in fig. 4, the splicing device includes: an elliptical image acquisition module 40, a reconstruction circle feature module 41 and a splicing matrix calculation module 42; wherein,
an elliptical image acquisition module 40, configured to acquire an elliptical image formed by the measurement positions of the planar circular target before and after movement;
a circle feature reconstruction module 41, configured to extract an ellipse from the photographed elliptical image of the planar circle target, fit an elliptical equation, and reconstruct circle features of the planar circle target in the coordinate system of the previous and subsequent local measurements;
and the splicing matrix calculation module 42 is used for constructing an optimized objective function according to the circle characteristics and solving the splicing matrix.
The circle features comprise center coordinates and normal vectors of a circle plane.
Specifically, the circle reconstruction feature module 41 may be configured to implement step 203; the splicing matrix calculation module 42 can be used to implement step 204, and is not described in detail here.
Further, the elliptical image acquisition module 40, before acquiring the elliptical image of the planar circular target at the measurement position before and after the movement, includes:
fixing and placing the measured object at a reasonable position which can be observed by a calibrated binocular vision system, and establishing a global coordinate system;
the planar circular target is placed in front of a measured object, and when the calibrated binocular vision system is moved, an elliptical image formed by the measuring positions of the planar circular target before and after the movement is shot.
The elliptical image acquisition module 40, the reconstruction circle characteristic module 41 and the splicing matrix calculation module 42 of the three-dimensional vision measurement data splicing device based on the plane circle target can be arranged in a binocular vision system.
In order to better illustrate the implementation effect of the splicing method, the planar circular target method can be compared with an ICP method and a planar grid target method respectively in an experiment, so as to evaluate the actual effect of the planar circular target method for three-dimensional data splicing. The experimental system is set up as shown in fig. 5, two AVT F302b type CCD cameras are adopted, the target surface size of each camera is 2/3inch, the resolution is 1028pixel multiplied by 960pixel, a binocular vision system is formed by combining two schnider 12mm lenses, and a three-dimensional scanning measuring head is formed by combining the binocular vision system with a projector.
a. Comparison of planar circular target method with ICP method:
the missile model is taken as a measured object, and the measured object is measured twice by using a three-dimensional scanning measuring head, as shown in fig. 6, fig. 6(a) and 6(b) are point cloud charts obtained at two measurement positions, wherein fig. 6(a) is position 1, and fig. 6(b) is position 2.
Fig. 7(a) is a point cloud image obtained by splicing by using the plane circle target method provided by the invention, and fig. 7(b) is a point cloud image obtained by splicing by using the ICP method, and as can be seen from fig. 7, the better splicing effect is obtained by splicing by using the plane circle target method, and the three-dimensional point cloud data obtained from two measurement positions are correctly spliced; by adopting the ICP method, splicing errors occur; in addition, the ICP method takes 213.6 seconds for splicing, whereas the planar circular target method of the present invention takes 6.8 seconds for splicing.
The reason why the ICP method has splicing error is that: the measured object has an insufficient surface texture and few features, and therefore, the ICP method is not suitable for measured objects having an insufficient surface texture, such as a cylindrical curved surface. As can be seen from comparison of the splicing experiment results, the method based on the planar circular target is superior to the ICP method in terms of splicing effect and algorithm processing time.
b. Comparing the plane circle target method with the plane grid target method:
under the condition of no shielding, splicing accuracies of a three-dimensional data splicing method based on a plane circular target in the directions of x, y and z axes are respectively 0.067mm, 0.035mm and 0.134mm under the experimental condition shown in FIG. 5; and calculating the camera coordinate system splicing matrix with the same position relation by using a plane grid target method to obtain splicing precisions of 0.057mm, 0.032mm and 0.103mm in the directions of the x axis, the y axis and the z axis respectively. Therefore, the experimental result shows that the precision of the plane circular target method provided by the invention is equivalent to that of the plane grid target method under the condition of no shielding.
In the case where the planar circular target is partially occluded, here, partial occlusion can be achieved by placing an occlusion object in front of the stitching target, and also under the experimental conditions as shown in fig. 5, the images taken by the right camera in the binocular vision system at the position 1 and the position 2 are shown in fig. 8, fig. 8(a) is the image taken by the right camera at the position 1, and fig. 8(b) is the image taken by the right camera at the position 2.
Under the condition that the planar circular target part is shielded, based on the planar grid target method, at two measurement positions (position 1 and position 2), the three-dimensional reconstruction result of the grid corner point is shown in fig. 9, where fig. 9(a) is a position 1 grid target three-dimensional point set, and fig. 9(b) is a position 2 grid target three-dimensional point set. Due to uncertainty of the shielding condition, there are various possibilities as shown in fig. 10(a) to 10(d) for matching between the reconstructed three-dimensional points, because shielding causes uncertainty of matching, and a planar grid target method causes that three-dimensional data cannot be correctly spliced; under the same condition, the method based on the plane circular target is not influenced by partial shielding, and the splicing precision in the directions of x, y and z axes is respectively 0.062mm, 0.037mm and 0.168 mm; experimental results show that the plane circular target method provided by the invention is superior to a plane grid target method under the condition of partial shielding, and is suitable for three-dimensional data splicing in a complex field environment.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention.
Claims (6)
1. A three-dimensional vision measurement data splicing method based on a plane circular target is characterized by comprising the following steps:
fixing and placing the measured object at a reasonable position which can be observed by a calibrated binocular vision system, and establishing a global coordinate system;
placing the planar circular target in front of the measured object, moving the calibrated binocular vision system, and shooting an elliptical image formed by the measuring positions of the planar circular target before and after movement;
extracting an ellipse from the shot plane circle target image, fitting an ellipse equation, and reconstructing circle characteristics of the plane circle target under the two-time local measurement coordinate system;
constructing an optimized objective function according to the circle characteristics, and solving a splicing matrix;
wherein the fitting ellipse equation is:
and fitting an elliptic equation by combining a direct least square method of a random sample consensus (RANSAC).
2. The method of claim 1, wherein the circular features comprise center coordinates and a normal vector to a circular plane.
3. The method according to claim 1, wherein the circle feature of the reconstructed planar circle target under the two previous and subsequent local measurement coordinate systems is:
and reconstructing the circle feature under the local measurement coordinate system by utilizing a geometrically intersected circle feature three-dimensional reconstruction method based on an elliptic equation, a camera parameter matrix and the radius value of the circle feature on the plane circle target which are fitted by combining a direct least square method of RANSAC.
4. The method of claim 1, wherein the constructing an optimization objective function from the circle features to solve the stitching matrix is:
constructing an optimization objective function by utilizing the identity of circle characteristics of the reconstructed front and rear measurement positions in space, and obtaining the ratio of inner points and outer points when fitting an elliptic equation according to RANSAC; and setting the weight corresponding to the matched edge point set, and calculating the high-precision splicing matrix by utilizing a Levenberg-Marquardt nonlinear optimization method.
5. A device for realizing three-dimensional vision measurement data splicing based on a planar circular target is characterized by comprising an elliptical image acquisition module, a circle reconstruction characteristic module and a splicing matrix calculation module; wherein,
the elliptical image acquisition module is used for acquiring elliptical images formed by measuring positions of the planar circular target before and after movement;
the circle feature reconstruction module is used for extracting an ellipse from the shot plane circle target ellipse image, fitting an ellipse equation and reconstructing circle features of the plane circle target in a local measurement coordinate system of the front and back times;
the splicing matrix calculation module is used for constructing an optimized objective function according to the circle characteristics and solving a splicing matrix;
wherein the fitting ellipse equation is:
an ellipse equation is fitted with a direct least squares fit in conjunction with RANSAC.
6. The apparatus of claim 5, wherein the elliptical image acquisition module, the circle reconstruction feature module, and the stitching matrix calculation module are disposed in a binocular vision system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210494953.9A CN102944191B (en) | 2012-11-28 | 2012-11-28 | Method and device for three-dimensional vision measurement data registration based on planar circle target |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210494953.9A CN102944191B (en) | 2012-11-28 | 2012-11-28 | Method and device for three-dimensional vision measurement data registration based on planar circle target |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102944191A CN102944191A (en) | 2013-02-27 |
| CN102944191B true CN102944191B (en) | 2015-06-10 |
Family
ID=47727156
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210494953.9A Expired - Fee Related CN102944191B (en) | 2012-11-28 | 2012-11-28 | Method and device for three-dimensional vision measurement data registration based on planar circle target |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102944191B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104154877B (en) * | 2014-09-03 | 2016-07-06 | 中国人民解放军国防科学技术大学 | A kind of complicated convex-surface type object dimensional is rebuild and volume measuring method |
| CN107690303B (en) * | 2015-06-04 | 2021-11-05 | 惠普发展公司，有限责任合伙企业 | Computing device and method for generating three-dimensional model |
| CN107192348A (en) * | 2016-03-14 | 2017-09-22 | 武汉小狮科技有限公司 | A kind of high-precision 3D vision measuring methods |
| CN106017321A (en) * | 2016-06-16 | 2016-10-12 | 沈阳飞机工业（集团）有限公司 | Binocular vision-based large-dimensional geometric quantity measurement method |
| CN106952344A (en) * | 2017-05-04 | 2017-07-14 | 西安新拓三维光测科技有限公司 | A kind of Damaged model calculation method for being used to remanufacture reparation |
| CN107220999A (en) * | 2017-06-19 | 2017-09-29 | 江南大学 | The research of workpiece circular arc Edge Feature Points matching process |
| CN108362220A (en) * | 2018-01-19 | 2018-08-03 | 中国科学技术大学 | The method of measuring three-dimensional morphology and defects detection for printed wiring board |
| CN114302173B (en) | 2021-12-31 | 2022-07-15 | 广东工业大学 | Two-dimensional image splicing system and method for planar coding target |
| CN114440776B (en) * | 2022-01-28 | 2024-07-19 | 上海交途科技有限公司 | Automatic displacement measurement method and system based on machine vision |
| CN115401536B (en) * | 2022-08-30 | 2024-04-12 | 深圳数马电子技术有限公司 | Reamer grinding method, reamer grinding device, numerical control machine, computer equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1566900A (en) * | 2003-06-11 | 2005-01-19 | 北京航空航天大学 | Vision measuring method for spaced round geometrical parameters |
| CN1971206A (en) * | 2006-12-20 | 2007-05-30 | 北京航空航天大学 | Calibration method for binocular vision sensor based on one-dimension target |
| CN101377404A (en) * | 2008-07-11 | 2009-03-04 | 北京航空航天大学 | Method for disambiguating space round gesture recognition ambiguity based on angle restriction |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001074428A (en) * | 1999-09-03 | 2001-03-23 | Sanyo Electric Co Ltd | Method and jig for calibrating shape measuring apparatus |
| FI112279B (en) * | 2001-11-21 | 2003-11-14 | Mapvision Oy Ltd | Method for determining offset points |
-
2012
- 2012-11-28 CN CN201210494953.9A patent/CN102944191B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1566900A (en) * | 2003-06-11 | 2005-01-19 | 北京航空航天大学 | Vision measuring method for spaced round geometrical parameters |
| CN1971206A (en) * | 2006-12-20 | 2007-05-30 | 北京航空航天大学 | Calibration method for binocular vision sensor based on one-dimension target |
| CN101377404A (en) * | 2008-07-11 | 2009-03-04 | 北京航空航天大学 | Method for disambiguating space round gesture recognition ambiguity based on angle restriction |
Non-Patent Citations (2)
| Title |
|---|
| 一种基于共面圆的摄像机自标定算法;赵征等;《光电子激光》;20091231;第20卷(第12期);正文第1-3节及图1-2 * |
| 基于平面基线靶标的视觉测量数据拼接方法;孙军华等;《机械工程学报》;20060731;第42卷(第7期);正文第2-3节 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102944191A (en) | 2013-02-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102944191B (en) | Method and device for three-dimensional vision measurement data registration based on planar circle target | |
| CN105205824B (en) | Multiple-camera global calibration method based on high-precision auxiliary camera and ball target | |
| CN105783775B (en) | A kind of minute surface and class minute surface object surface appearance measuring device and method | |
| CN103759670B (en) | A kind of object dimensional information getting method based on numeral up short | |
| Scaramuzza et al. | A flexible technique for accurate omnidirectional camera calibration and structure from motion | |
| CN104596502B (en) | Object posture measuring method based on CAD model and monocular vision | |
| Zhang | Camera calibration with one-dimensional objects | |
| CN103162622B (en) | The Portable ball target of single camera vision system and use thereof and measuring method thereof | |
| CN103714571B (en) | A kind of based on photogrammetric single camera three-dimensional rebuilding method | |
| CN110672039A (en) | Object omnibearing three-dimensional measurement method based on plane reflector | |
| CN105551039A (en) | Calibration method and calibration device for structured light 3D scanning system | |
| CN104990515A (en) | Three-dimensional shape measurement system and method for large-size object | |
| Albarelli et al. | Robust camera calibration using inaccurate targets | |
| Hu et al. | A four-camera videogrammetric system for 3-D motion measurement of deformable object | |
| CN104766292A (en) | Method and system for calibrating multiple stereo cameras | |
| CN102810205A (en) | Method for calibrating camera shooting or photographing device | |
| CN108535097A (en) | A kind of method of triaxial test sample cylindrical distortion measurement of full field | |
| CN105547190B (en) | 3 D measuring method and device based on double angle unifrequency fringe projections | |
| CN106705860B (en) | A kind of laser distance measurement method | |
| CN107038753B (en) | Stereoscopic vision three-dimensional reconstruction system and method | |
| Jiang et al. | Combined shape measurement based on locating and tracking of an optical scanner | |
| CN109087355A (en) | The monocular camera pose measuring apparatus and method updated based on iteration | |
| Yang et al. | Accurate calibration approach for non-overlapping multi-camera system | |
| Daftry et al. | Flexible and User-Centric Camera Calibration using Planar Fiducial Markers. | |
| Fernandez et al. | Planar-based camera-projector calibration |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20150610 |