CN111780689A - Optimal rotation angle determination method based on cross-correlation structured light 360-degree measurement - Google Patents
Optimal rotation angle determination method based on cross-correlation structured light 360-degree measurement Download PDFInfo
- Publication number
- CN111780689A CN111780689A CN202010669239.3A CN202010669239A CN111780689A CN 111780689 A CN111780689 A CN 111780689A CN 202010669239 A CN202010669239 A CN 202010669239A CN 111780689 A CN111780689 A CN 111780689A
- Authority
- CN
- China
- Prior art keywords
- measurement
- cross
- rotation angle
- correlation
- determining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
- G01B11/25—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures by projecting a pattern, e.g. one or more lines, moiré fringes on the object
- G01B11/2518—Projection by scanning of the object
Abstract
The invention provides an optimal rotation angle judgment method for 360-degree measurement of structured light based on cross correlation, and belongs to the field of three-dimensional measurement of structured light. The method comprises the five steps of determining the rotation times, setting the angles of the two cameras, establishing a cross-correlation curve, determining the optimal rotation angle and finishing the reconstruction of 360 degrees. The method of the invention utilizes the prior knowledge of roughly estimated object external ellipsoid to set the measuring times needed by the 360-degree reconstruction of the object to be measured; taking an image acquired by a left camera in the previous measurement as a reference, performing cross-correlation operation with an image acquired by a right camera in real time in the rotating process of the turntable, establishing a cross-correlation curve, searching a curve peak value, and determining a rotating angle corresponding to the peak value as the optimal rotating angle of the turntable; stopping rotating the rotary table at the position from the cross-correlation curve to the peak value, and carrying out next measurement; and repeating the steps to finally finish the 360-degree measurement of the object.
Description
Technical Field
The invention relates to the technical field of structured light three-dimensional measurement, in particular to an optimal rotation angle judgment method based on cross-correlation structured light 360-degree measurement.
Background
The structured light 360-degree measurement technology is a measurement method for acquiring 360-degree three-dimensional information of an object to be measured by fusing the structured light measurement technology with a three-dimensional point cloud matching technology. In the structured light measurement technology, due to the limitation of the field range of the system formed by the camera and the projector, only part of the information of the object surface can be acquired in a single measurement. In order to obtain 360-degree information of the surface of the object, the object needs to be rotated in the measuring process, so that the part of the object, which is positioned in the measuring blind area, rotates into the measurable area. However, the object rotation may cause the reconstructed point clouds not to be in the same object coordinate system, so that the point cloud matching technique is required to remove the difference between the reconstructed coordinate systems introduced by the object motion. Meanwhile, due to the influence of the surface structure of the object, objects with different shapes have the optimal measurement times, so that the optimal rotation angle exists during measurement. The existing method mostly adopts a precise rotating table to control the rotating angle, and the problem of the optimal rotating angle is not considered yet. If the same measurement times are adopted for objects with different complexity degrees, partial visual angles of the complex objects are lost, redundant information of point cloud cavities or simple objects is excessive, and the point cloud fusion difficulty is increased. And the angle is controlled by using the precise rotating table, so that the dependence degree of the equipment on a precise instrument is increased.
Therefore, it is necessary to provide a new optimal rotation angle determination method based on the 360 ° measurement of the correlated structured light to solve the above technical problems.
Disclosure of Invention
The invention aims to provide an optimal rotation angle judgment method based on cross-correlation structured light 360-degree measurement, which solves the problem of angle judgment in the traditional structured light 360-degree three-dimensional measurement method without controlling the rotation angle through a precision rotating table.
In order to solve the technical problem, the method for determining the optimal rotation angle based on the 360-degree measurement of the cross-correlation structured light provided by the invention comprises the following steps:
s1: determining the number of revolutions
The method for determining the rotation times is used in measurement and is obtained according to the maximum part of the curvature of an ellipsoid circumscribed by an object; according to the method, a left camera internal reference matrix is firstly calibrated, wherein the left camera internal reference matrix comprises a camera focal length f, a pixel size and a focal plane principal point position, an external rectangular body is roughly estimated along the long side and the short side of an object, and an optical center d from the center of a rotary table to a left camera of the camera; and an external rectangular body roughly estimates the ellipsoid of the contained object according to the formula (1) as follows:
wherein, a, b and c are the length, width and height of the circumscribed rectangle of the object.
During modeling, the short axis of the ellipsoid directly faces the left camera as a starting point, and a relation between the rotation angle and the image overlap ratio acquired by the camera is established; when the coincidence ratio of the previous measurement and the subsequent measurement is 1/2, the rotation angle is the optimal rotation angle theta; acquiring an Image with a short axis aligned with the left camera by using a feature matching methodrefImage at the time of later rotationcapMatching in real time, when Image is matchedrefAnd ImagecapThe feature point matching points in (1) are respectively located in ImagerefLeft side and ImagecapAnd on the right side, the current angle is the required optimal rotation angle theta.
The number of rotations required to measure the object is:
s2: setting dual camera angles
The invention ensures that the rotation angle is the optimal rotation angle every time through the included angle between the two cameras and the later algorithm, and firstly, the included angle between the two cameras is set as theta; acquiring an external parameter matrix R, T between the two cameras by using the existing calibration method, and setting an included angle between the two cameras as theta by observing the value of R; wherein the relationship between R and θ is shown in equation (3):
where tr denotes the trace of the matrix.
S3: establishing a cross-correlation curve
After the current measurement is completed, the object needs to be rotated by theta for the next measurement. In the rotating process, performing cross-correlation operation on an image acquired by a currently measured left camera and an image acquired by a right camera in real time when the image is rotated to a next measuring position to establish a cross-correlation curve based on time;
calculating the position of the gravity center coordinate of the image according to a formula (4) for the image acquired by the left camera and the image acquired by the right camera in real time;
and calculating the abscissa distance and the ordinate distance between the object point contour point farthest from the center of gravity and the center of gravity by taking the center of gravity as the center, and taking the circumscribed rectangle of the two-dimensional projection contour of the object. Aligning the gravity centers of the images collected by the left camera and the right camera, and performing cross-correlation operation on the circumscribed rectangle part of the image collected by the right camera and the image collected by the left camera as shown in a formula (5):
wherein, AvgLAnd AvgRRespectively representing the average value of the gray scales in the circumscribed rectangle of the images collected by the left camera and the right camera,2(IL) And2(IR) Respectively representing the gray standard deviation in a set field on one image;
because the included angle of theta exists on the horizontal square of the left camera and the right camera, N is changed into a quadratic convex function along with time, and a curve which is drawn by taking t as a variable is a cross-correlation curve. S4: determining an optimal rotation angle
Fitting a function by using the cross-correlation function curve obtained in the step S3, calculating first derivatives of functions on the left side and the right side of the previous point of the current point in real time, if the previous point is a function peak value, stopping rotating the turntable immediately, and determining the position as the optimal rotation angle of the measurement without the assistance of a precision instrument; the manner of determining the peak point of the function is shown in equation (6):
wherein theta isnI.e. the rotation angle at which the peak point occurs.
S5: complete 360 deg. reconstruction
After the rotating table completes 360-degree rotation, the 360-degree reconstruction of the object can be completed completely by using the existing point cloud splicing method.
Compared with the related art, the optimal rotation angle determination method based on the cross-correlation structured light 360-degree measurement has the following beneficial effects:
the invention provides an optimal rotation angle judgment method based on cross-correlation structured light 360-degree measurement, which determines the optimal rotation angle of an object in 360-degree measurement according to the maximum curvature part of an ellipsoid circumscribed by the object, thereby determining the measurement times; the problems of point cloud redundancy, point cloud cavities and the like caused by excessive measurement times or small measurement times during measurement are reduced; the rotation angle of the object is ensured to be the optimal rotation angle required by measurement in each measurement by utilizing a cross-correlation matching algorithm through the geometric relationship between the two cameras; the dependence of the system on a precise stepping motor in measurement is removed.
Drawings
FIG. 1 is a flow chart of the process of the present invention;
FIG. 2 is a system diagram of the method of the present invention;
FIG. 3 is a relationship between a cross-correlation curve and a rotation angle;
fig. 4 is a graph of 360 ° measurements of an object, (a) the reconstruction is rotated by 0 °, and (b) the reconstruction is rotated by 160 °.
The reference numbers in the figures are:
1. cameras A, 2, a projector 3, cameras B, 4, a computer 5, a rotating table 6, an object to be measured 7 and ILAnd at each angle IRCross correlation value of 8, I at each angle R9 at initial angle IL。
Detailed Description
The invention is further described with reference to the following figures and embodiments.
The following describes an exemplary embodiment of an optimal rotation angle determination method based on cross-correlation structured light 360 ° measurement according to the present invention in detail, and the present invention is further described in detail. It should be noted that the following examples are only for illustrative purposes and should not be construed as limiting the scope of the present invention, and that the skilled person in the art may make modifications and adaptations of the present invention without departing from the scope of the present invention. Preferably, the present example uses a conventional rotating platform to move the object for 360 ° reconstruction, and the system apparatus is shown in fig. 2. The specific implementation steps are as follows:
s1: determining the number of revolutions
The method for determining the rotation times used in the measurement is obtained according to the maximum curvature part of the circumscribed ellipsoid of the object. According to the method, a left camera internal reference matrix is firstly calibrated, wherein the left camera internal reference matrix comprises a camera focal length f, a pixel size and a focal plane principal point position, an external rectangular body is roughly estimated along the long side and the short side of an object, and an optical center d from the center of a rotary table to a left camera of the camera. And an external rectangular body roughly estimates the ellipsoid of the contained object according to the formula (1) as follows:
wherein, a, b and c are the length, width and height of the circumscribed rectangle of the object.
And (3) taking the short axis of the ellipsoid body directly opposite to the left camera as a starting point, and establishing a relation between the rotation angle and the coincidence degree of the images acquired by the camera. When the coincidence ratio of the previous measurement and the subsequent measurement is 1/2, the rotation angle is the optimum rotation angle θ. Acquiring an Image with a short axis aligned with the left camera by using a feature matching methodrefImage at the time of later rotationcapMatching in real time, when Image is matchedrefAnd ImagecapThe feature point matching points in (1) are respectively located in ImagerefLeft side and ImagecapAnd on the right side, the current angle is the required optimal rotation angle theta. The number of rotations required to measure the object is:
s2: setting dual camera angles
The invention ensures that the rotation angle is the optimal rotation angle every time through the included angle between the two cameras and the later algorithm, and firstly, the included angle between the two cameras is set as theta. By utilizing the existing calibration method, the external parameter matrix R, T between the two cameras is obtained, and the included angle between the two cameras can be set as theta by observing the value of R. Wherein the relationship between R and θ is shown in equation (3):
where tr denotes the trace of the matrix.
S3: establishing a cross-correlation curve
After the current measurement is completed, the object needs to be rotated by theta for the next measurement. In the rotating process, the image acquired by the left camera which is currently measured and the image acquired by the right camera in real time when the left camera rotates to the next measuring position are subjected to cross-correlation operation, and a cross-correlation curve based on time is established.
And (3) calculating the gravity center coordinate position of the image according to a formula (4) for the image acquired by the left camera and the image acquired by the right camera in real time:
and calculating the abscissa distance and the ordinate distance between the object point contour point farthest from the center of gravity and the center of gravity by taking the center of gravity as the center, and taking the circumscribed rectangle of the two-dimensional projection contour of the object. Aligning the gravity centers of the images collected by the left camera and the right camera, and performing cross-correlation operation on the circumscribed rectangle part of the image collected by the right camera and the image collected by the left camera as shown in a formula (5):
wherein, AvgLAnd AvgRRespectively representing the average value of the gray scales in the circumscribed rectangle of the images collected by the left camera and the right camera,2(IL) And2(IR) The gray scale standard deviations in the set fields on one image are respectively represented.
Because the included angle of theta exists on the horizontal square of the left camera and the right camera, N is changed into a quadratic convex function along with time, and a curve which is drawn by taking t as a variable is a cross-correlation curve.
S4: determining an optimal rotation angle
And fitting the function by using the cross-correlation function curve obtained in the step S3, calculating first derivatives of functions on the left side and the right side of the previous point of the current point in real time, and if the previous point is the function peak value, stopping rotating the turntable immediately, wherein the position is the optimal rotation angle of the next measurement without the assistance of a precision instrument. The manner of determining the peak point of the function is shown in equation (6):
wherein theta isnI.e. the rotation angle at which the peak point occurs.
S5: complete 360 deg. reconstruction
After the rotating table completes 360-degree rotation, the 360-degree reconstruction of the object can be completed completely by using the existing point cloud splicing method.
The working principle of the optimal rotation angle judgment method based on the 360-degree measurement of the cross-correlation structured light provided by the invention is as follows:
the rotation angle is not required to be controlled by a precise rotating table, before measurement, the measurement times are divided according to the measurement complexity of an object, and an approximate included angle between the two cameras is selected according to the measurement times; after the included angle of the two cameras is determined, the included angle between the two cameras is accurately calibrated, the image acquired by the left camera in the previous time and the image acquired by the right camera in real time are used for performing cross-correlation operation in the measurement, the rotation angle of the object is monitored in real time, and whether a reasonable area is processed or not is judged; after a cross-correlation matched line is obtained, detecting a curve peak value in real time, determining an optimal rotating position, and carrying out next measurement on the object; and repeating the steps to finally complete the 360-degree reconstruction of the object.
The above description is only an embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (9)
1. An optimal rotation angle determination method based on cross-correlation structured light 360-degree measurement is characterized by comprising the following steps:
s1: determining the number of rotations;
s2: setting angles of the two cameras;
s3: establishing a cross-correlation curve;
s4: determining an optimal rotation angle;
s5: the 360 deg. reconstruction is completed.
2. The method for determining an optimal rotation angle based on 360 ° measurement of structured light with cross-correlation according to claim 1, wherein the step S1 specifically includes the following steps: roughly estimating an external ellipsoid of the object, modeling by using the ellipsoid and the relative pose of the ellipsoid and a camera, and determining the rotation times of the object during measurement through the maximum curvature part of the external ellipsoid.
3. The method for determining an optimal rotation angle based on 360 ° measurement of structured light with cross-correlation according to claim 1, wherein the step S2 specifically includes the following steps: and calibrating an included angle theta between the two cameras by using the rotation times.
4. The method for determining an optimal rotation angle based on 360 ° measurement of structured light with cross-correlation according to claim 1, wherein the step S3 specifically includes the following steps: and taking the left camera as a reference, and carrying out real-time cross-correlation operation on an object image acquired by the left camera in the previous measurement and an object image acquired by the right camera to establish a cross-correlation curve.
5. The method for determining an optimal rotation angle based on 360 ° measurement of structured light with cross-correlation according to claim 1, wherein the step S4 specifically includes the following steps: and establishing a cross-correlation curve in real time, and when the curve reaches the peak position, stopping rotating the rotary table, and calculating rigid body transformation parameters of the object at the moment for the measured optimal rotation angle.
6. The method for determining an optimal rotation angle based on 360 ° measurement of structured light with cross-correlation according to claim 1, wherein the step S5 specifically includes the following steps: and unifying three-dimensional point clouds obtained by multiple measurements to the same object coordinate system through the transformation parameters between adjacent measurements to complete the 360-degree reconstruction of the object to be measured.
7. The method for determining an optimal rotation angle based on 360 ° measurement of structured light with cross correlation as claimed in claim 1, wherein in step S1, the ellipsoid containing the object is roughly estimated according to the circumscribed rectangle of the object as:
wherein, a, b and c are the length, width and height of the circumscribed rectangle of the object; during modeling, the short axis of the ellipsoid directly faces the left camera as a starting point, and a relation between the rotation angle and the image overlap ratio acquired by the camera is established; when the coincidence ratio of the previous measurement and the subsequent measurement is 1/2, the rotation angle is the optimal rotation angle theta; the number of measurements is then:
8. the method for determining an optimal rotation angle based on 360 ° measurement of structured light with cross correlation according to claim 1, wherein in step S2, an extrinsic parameter matrix R, T between the two cameras is obtained, and an included angle between the two cameras is set to θ by observing a value of R; wherein the relationship between R and θ is:
where tr denotes the trace of the matrix.
9. The method for determining an optimal rotation angle based on 360 ° measurement of correlated structured light according to claim 1, wherein in step S3, the image I captured by the left camera is measured currentlyLAnd an image I acquired by the right camera in real time when the right camera rotates to the next measurement positionRPerforming cross-correlation operation, and establishing a cross-correlation curve based on time;
wherein, AvgLAnd AvgRRespectively representing the average gray level in the circumscribed rectangle of the images collected by the left and right cameras, (X)LC,YLC) And (X)RC,YRC) The barycentric coordinates of the object regions of the images collected by the left camera and the right camera respectively, t is a time function,2(IL) And2(IR) Respectively representing the gray standard deviation in a set field on one image; when the curve has a peak value, the rotating disc stops rotating, and the position is the optimal rotating angle of the measurement.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010669239.3A CN111780689B (en) | 2020-07-13 | 2020-07-13 | Optimal rotation angle determination method based on cross-correlation structured light 360-degree measurement |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010669239.3A CN111780689B (en) | 2020-07-13 | 2020-07-13 | Optimal rotation angle determination method based on cross-correlation structured light 360-degree measurement |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111780689A true CN111780689A (en) | 2020-10-16 |
| CN111780689B CN111780689B (en) | 2021-06-29 |
Family
ID=72768163
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010669239.3A Active CN111780689B (en) | 2020-07-13 | 2020-07-13 | Optimal rotation angle determination method based on cross-correlation structured light 360-degree measurement |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111780689B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103279987A (en) * | 2013-06-18 | 2013-09-04 | 厦门理工学院 | Object fast three-dimensional modeling method based on Kinect |
| US20140104105A1 (en) * | 2010-07-30 | 2014-04-17 | Donald C.D. Chang | Polarization Re-alignment for Mobile Satellite Terminals |
| CN103940369A (en) * | 2014-04-09 | 2014-07-23 | 大连理工大学 | Quick morphology vision measuring method in multi-laser synergic scanning mode |
| CN106500619A (en) * | 2016-10-21 | 2017-03-15 | 哈尔滨理工大学 | The camera internal imageing sensor alignment error separation method of view-based access control model measurement |
| CN107490346A (en) * | 2017-08-17 | 2017-12-19 | 江苏省质量和标准化研究院 | A kind of RFID multi-tags Network Three-dimensional measurement modeling method of view-based access control model |
| CN107702663A (en) * | 2017-09-29 | 2018-02-16 | 五邑大学 | A kind of point cloud registration method based on the rotation platform with index point |
| CN111189416A (en) * | 2020-01-13 | 2020-05-22 | 四川大学 | Structural light 360-degree three-dimensional surface shape measuring method based on characteristic phase constraint |
-
2020
- 2020-07-13 CN CN202010669239.3A patent/CN111780689B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140104105A1 (en) * | 2010-07-30 | 2014-04-17 | Donald C.D. Chang | Polarization Re-alignment for Mobile Satellite Terminals |
| CN103279987A (en) * | 2013-06-18 | 2013-09-04 | 厦门理工学院 | Object fast three-dimensional modeling method based on Kinect |
| CN103940369A (en) * | 2014-04-09 | 2014-07-23 | 大连理工大学 | Quick morphology vision measuring method in multi-laser synergic scanning mode |
| CN106500619A (en) * | 2016-10-21 | 2017-03-15 | 哈尔滨理工大学 | The camera internal imageing sensor alignment error separation method of view-based access control model measurement |
| CN107490346A (en) * | 2017-08-17 | 2017-12-19 | 江苏省质量和标准化研究院 | A kind of RFID multi-tags Network Three-dimensional measurement modeling method of view-based access control model |
| CN107702663A (en) * | 2017-09-29 | 2018-02-16 | 五邑大学 | A kind of point cloud registration method based on the rotation platform with index point |
| CN111189416A (en) * | 2020-01-13 | 2020-05-22 | 四川大学 | Structural light 360-degree three-dimensional surface shape measuring method based on characteristic phase constraint |
Non-Patent Citations (2)
| Title |
|---|
| XIAOHUI QIU: "An Non-Parametric Rotating Angle Acquisition Method for Optimal ISAR Imaging", 《ANTENNAS AND PROPAGATION SOCIETY INTERNATIONAL SYMPOSIUM 2006, IEEE》 * |
| 王云舒: "特殊视觉应用环境下微小型飞行器视觉_惯性组合导航关键技术", 《中国博士学位论文全文数据库 信息科技辑》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111780689B (en) | 2021-06-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108921901B (en) | Large-view-field camera calibration method based on precise two-axis turntable and laser tracker | |
| US8363979B2 (en) | Method for ascertaining the axis of rotation of a vehicle wheel | |
| CN109377551B (en) | Three-dimensional face reconstruction method and device and storage medium thereof | |
| CN109579695B (en) | Part measuring method based on heterogeneous stereoscopic vision | |
| CN111091599B (en) | Multi-camera-projector system calibration method based on sphere calibration object | |
| CN104019745B (en) | Based on the free planar dimension measuring method of single visual feel indirect calibration method | |
| CN104634248B (en) | Revolving shaft calibration method under binocular vision | |
| CN109859272A (en) | A kind of auto-focusing binocular camera scaling method and device | |
| CN103559735A (en) | Three-dimensional reconstruction method and system | |
| CN109272555B (en) | External parameter obtaining and calibrating method for RGB-D camera | |
| CN103759669A (en) | Monocular vision measuring method for large parts | |
| CN109191527B (en) | Alignment method and device based on minimum distance deviation | |
| CN113724337B (en) | Camera dynamic external parameter calibration method and device without depending on tripod head angle | |
| Chen et al. | A self-recalibration method based on scale-invariant registration for structured light measurement systems | |
| CN116295113A (en) | Polarization three-dimensional imaging method integrating fringe projection | |
| CN110044266B (en) | Photogrammetry system based on speckle projection | |
| CN115187612A (en) | Plane area measuring method, device and system based on machine vision | |
| CN108257184B (en) | Camera attitude measurement method based on square lattice cooperative target | |
| CN111780689B (en) | Optimal rotation angle determination method based on cross-correlation structured light 360-degree measurement | |
| CN103559710B (en) | A kind of scaling method for three-dimensional reconstruction system | |
| CN108682035B (en) | Laser space plane equation calculation method of single-step optical knife system | |
| CN115046497A (en) | Improved calibration method based on grating projection measurement system | |
| CN114170321A (en) | Camera self-calibration method and system based on distance measurement | |
| Zhang et al. | Camera Calibration Algorithm for Long Distance Binocular Measurement | |
| CN114485386A (en) | Method, device and system for calibrating workpiece coordinate system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |