CN105627917A - Large-scale structural component assembly joining measurement method based on visual principle - Google Patents

Abstract

The invention belongs to the field of mechanical assembly manufacturing technology, in particular relates to large-scale structural component assembly joining measurement method based on the visual principle, and aims to solve the problems of either high cost or large errors of conventional large-scale structural component assembly joining measurement. The method comprises the steps of setting and calibrating a measurement camera, mounting optical control points on structural components to be measured, measuring three-dimensional coordinates of the optical control points, establishing an assembly coordinate system and solving the attitude deviation among the assembled structural components. The method solves the problems of online measurement of the six-degree of freedom attitude deviation in the large-scale structural component assembly joining process by using the stereoscopic visual principle of a dual camera. The method is simple in equipment measurement, low in cost, and capable of detecting elastic deformation of large-scale structural components. Experimental results show that when the assembly measurement space is 5m*5m*5m and the number of adopted feature light spots reaches 5, the root-mean-square error of the attitude measurement method maintains within 0.05 degrees, the refresh rate reaches 200 frames per second, and the data are accurate and reliable.

Description

A kind of large-sized structural parts assembling docking measuring method of view-based access control model principle

Technical field

The invention belongs to mechanical assembly manufacturing technology field, be specifically related to the large-sized structural parts assembling docking measuring method of a kind of view-based access control model principle.

Background technology

In the large-scale processing and manufacturing field such as aircraft, boats and ships, the accurate assembly of Large Scale Mechanical Structure is always up the technical barrier in production process. In the assembling of aircraft wing and fuselage, owing to physical dimension is big, interface scope is big, cannot judge whether wing attitude adjusts by eye-observation and align to fuselage, and assembling process very easily causes the damage of fuselage and inefficiency. Therefore need two body relative attitude measurement in assembling process, it is ensured that assembling process is quick and precisely. Measurement technology in external large-sized structural parts assembling docking, mainly has laser tracker measuring method and IndoorGPS measuring method at present.

Tracker measuring method (laser tracker for Leica company) is the measurement scheme utilizing laser tracker to coordinate its adnexa T-MAC, and laser tracker can measure the 3 d pose of T-MAC. Two T-MAC are separately fixed on to be assembled two structural member, one T-MAC of every tracker real-time tracking, spatial relationship between two tracker coordinate systems of calibration in advance, by under the 3 d pose unification of fixture and movable part to world coordinate system, and the movement executing mechanism assembling docking is instructed by the attitude data of two structural members, complete assembling process. But the method can only measure the attitude near T-MAC installation site, when large-sized structural parts, due to the elastic deformation of structural member, the method can bring bigger measurement error.

IndoorGPS measures system and is mainly made up of emitter and receptor. Arrange no less than 4 emitters in measurement space, fixture and movable part are respectively provided with no less than 4 receptors, the spatial relationship between each receptor of calibration in advance and tested section coordinate system. The three-dimensional coordinate of the real-time measuring receiver of system, using receptor as control point, resolves the relative dimensional attitude relation between tested assembling structural member. But the domestic IndoorGPS product that there is presently no maturation, equipment relies on import, relatively costly.

Summary of the invention

Present invention aim to address the problem that the assembling docking of existing large-sized structural parts is measured or cost is high or error is big, it is provided that the large-sized structural parts assembling docking measuring method of a kind of view-based access control model principle.

The present invention is achieved in that

The large-sized structural parts assembling docking measuring method of a kind of view-based access control model principle, comprises the steps:

The first step: arrange that measuring camera goes forward side by side rower calmly;

Second step: install optics control point on tested structural member;

3rd step: measure the three-dimensional coordinate at optics control point;

4th step: set up assembling coordinate system;

5th step: the attitude misalignment between real-time resolving assembling structural member.

As above layout is measured camera and carries out demarcating steps, two structural members to be assembled are placed in assembly shop and adjust in unwheeling, defining one of them structural member is fixture, in assembling process, position and attitude does not all change, another structural member is movable part, by adjusting position and the attitude of movable part so that it is alignment fixture completes assembling; Arrange two measurement cameras, it is desirable to two camera intersections are arranged, intersection angle about about 60 ��, and make intersection visual field cover two structural member assembly spaces; After fixed camera, the inner parameter of camera and external parameter are demarcated, so as to constitute double camera to measure system.

Installation optics control point step on tested structural member as above, fixes several optics control point respectively, it is simple to vision measurement system carries out feature recognition and extraction on two tested structural members to be assembled; Above-mentioned optics control point dispersed placement, covered structure part profile.

The three-dimensional coordinate step measuring optics control point as above, constitutes double camera Stereo Vision Measurement System, obtains the location parameter between the distortion parameter of camera and two cameras after camera calibration.

The three-dimensional coordinate step measuring optics control point as above, structural member is shot by double camera simultaneously, and left and right camera carries out image procossing and feature extraction, obtains the image coordinate of optical characteristic point, i.e. (x_L,y_L) and (x_L,y_L), resolve all optics control point 3 d space coordinate (X, Y, Z) under camera coordinates system according to equation below;

$\left\{\begin{array}{c}X={\mathrm{Zx}}_L/f_L\\ Y={\mathrm{Zy}}_L/f_L\\ Z=\frac{f_L(f_R{t}_x-x_L{t}_z)}{x_R(r_7{x}_L+r_8{y}_L+r_9{f}_L)-f_R(r_1{x}_L+r_2{y}_L+r_3{f}_L)}\\ =\frac{f_L(f_R{t}_y-y_L{t}_z)}{y_R(r_7{x}_L+r_8{y}_L+r_9{f}_L)-f_R(r_4{x}_L+r_5{y}_L+r_6{f}_L)}\end{array}\right.---\left(1\right)$

In formula: left camera coordinates system O_L-X_LY_LZ_LFor world coordinate system, right camera coordinates system O_R-X_RY_RZ_RRepresent relative to the spatial relationship spin matrix R and translation matrix T of world coordinate system, (x_L,y_L) and (x_L,y_L) picpointed coordinate of respectively left and right image; f_L��f_RRepresent the effective focal length of left and right camera respectively; f_L��f_RRepresent the effective focal length of left and right camera respectively; Matrix R_CameraAnd T_CameraFor:

As above foundation assembles coordinate system step, second step m fixture characteristic point of acquisition and n movable part characteristic point the three-dimensional coordinate on designing a model after having assembled; This m+n some three-dimensional coordinate under camera coordinates system is measured by the 3rd pacing; By m the fixture characteristic point coordinate under Two coordinate system respectively, utilize Coordinate Conversion principle, resolve the transformational relation tried to achieve between assembling coordinate system and camera coordinates system, if transition matrix is R_Coordinate; Utilize transition matrix R_CoordinateBe transformed into by n movable part characteristic point three-dimensional coordinate under camera coordinates system under assembling coordinate system, then now the three-dimensional coordinate of all characteristic points is all unified under assembling coordinate system.

Attitude misalignment step between real-time resolving as above assembling structural member, in assembling process, by double camera vision measurement system, the characteristic point on structural member being measured in real time, the movement parts characteristic point recorded in each moment coordinate under assembling coordinate system is as P_i, movement parts characteristic point coordinate under trim designs model is as P_i', use equation below to calculate spin matrix R and translation matrix T:

P��_i=R P_i+T(2)

Utilize spin matrix and translation matrix, it is achieved attitude algorithm, export six degree of freedom assembling deviation ��, beta, gamma, X_t��

Attitude misalignment step between real-time resolving as above assembling structural member,

$R=\left[\begin{array}{ccc}\cos \mathrm{\β}\cos \mathrm{\γ}& -\cos \mathrm{\α}\sin \mathrm{\γ}+\sin \mathrm{\α}\sin \mathrm{\β}\cos \mathrm{\γ}& \sin \mathrm{\α}\sin \mathrm{\γ}+\cos \mathrm{\α}\sin \mathrm{\β}\cos \mathrm{\γ}\\ \cos \mathrm{\β}\sin \mathrm{\γ}& \cos \mathrm{\α}\cos \mathrm{\γ}+\sin \mathrm{\α}\sin \mathrm{\β}\sin \mathrm{\γ}& -\sin \mathrm{\α}\cos \mathrm{\γ}+\cos \mathrm{\α}\sin \mathrm{\β}\sin \mathrm{\γ}\\ -\sin \mathrm{\β}& \sin \mathrm{\α}\cos \mathrm{\β}& \cos \mathrm{\α}\cos \mathrm{\β}\end{array}\right];$

Wherein ��, beta, gamma is respectively around x, y, the anglec of rotation of z-axis;

$T=\left[\begin{array}{c}{X}_t\\ Y_t\\ Z_t\end{array}\right];$

Wherein X_t, Y_t, Z_tRespectively along x, y, the translational movement of z-axis.

The invention has the beneficial effects as follows:

The present invention adopts and arranges and measure that camera is gone forward side by side that rower is fixed, installed optics control point on tested structural member, the three-dimensional coordinate of measuring optics control point, the attitude misalignment set up between assembling coordinate system and real-time resolving assembling structural member set step, adopt double camera principle of stereoscopic vision, solve the six-degree-of-freedom posture deviation on-line measurement problem in large-sized structural parts assembling docking operation, the method measurement device is simple, cost is low, and can detect the elastic deformation of large-sized structural parts. Experimental result shows, when alignment measurement space is 5m �� 5m �� 5m, and the feature luminous point of employing reaches 5, the root-mean-square error of this attitude measurement method is maintained within 0.05 ��, and refresh rate reaches 200 frames/second, and data are accurately and reliably.

Accompanying drawing explanation

Fig. 1 is the flow chart of the large-sized structural parts assembling docking measuring method of a kind of view-based access control model principle of the present invention;

Fig. 2 arranges in Fig. 1 that measuring camera goes forward side by side the fixed schematic diagram of rower;

In figure, 1-assembles coordinate system; 2-fixed structure piece; 3-optical characteristic point; 4-bascule part; 5-equipment adjustment unwheeling; 6-metering computer; The right camera of 7-; The left camera of 8-; 9-camera coordinates system.

Detailed description of the invention

Below in conjunction with drawings and Examples, the large-sized structural parts of a kind of view-based access control model principle of the present invention is assembled docking measuring method to be introduced:

As it is shown in figure 1, the large-sized structural parts assembling docking measuring method of a kind of view-based access control model principle, comprise the steps:

The first step: arrange that measuring camera goes forward side by side rower calmly;

Two structural members to be assembled are placed in assembly shop and adjust in unwheeling, defining one of them structural member is fixture, and in assembling process, position and attitude does not all change, and another structural member is movable part, by adjusting position and the attitude of movable part so that it is alignment fixture completes assembling. According to assembly shop's field condition, arrange two measurement cameras, it is desirable to two camera intersections are arranged, intersection angle about about 60 ��, and make intersection visual field cover two structural member assembly spaces. After fixed camera, the inner parameter of camera and external parameter are demarcated, so as to constitute double camera measure system, the method for demarcation can select tradition scaling method, based on any means such as the scaling method of active vision, self-calibrating methods. After demarcation, the parameters such as the position of camera, direction, focal length all can not change again.

In the present embodiment, as shown in Figure 2, structural member to be assembled is installed in equipment adjustment unwheeling 5, it is respectively defined as fixed structure piece 2 and bascule part 4, fixed structure piece 2 position and attitude in whole assembling process does not all change, equipment adjustment unwheeling 5 is relied on to adjust the position and attitude alignment fixed structure piece 2 of bascule part 4, it is achieved assembling. Fixing more than 3 optical characteristic point 3 on structural member to be assembled respectively, its theory three-dimensional coordinate under assembling coordinate system 1 is known. Setting up left camera 8 and right camera 7 so that two cameras all cover whole assembly space, constitute Stereo Vision Measurement System after calibration for cameras, setting up camera coordinates is 9. Metering computer is used for realizing image acquisition and data resolving.

Second step: install optics control point on tested structural member;

Two tested structural members to be assembled fix several optics control point respectively, it is simple to vision measurement system carries out feature recognition and extraction. Above-mentioned optics control point dispersed placement, covered structure part profile. In the present embodiment, require that characteristic point coordinate position under assembling coordinate system is known, realize typically via design reservation feature point installing hole on structural member, it is also possible to after characteristic point is installed, obtain, by externally measured method, the coordinate relation that characteristic point is docked between benchmark with assembling. Assume m characteristic point P of installation on fixture₁(x₁,y₁,z₁),P₂(x₂,y₂,z₂)...P_m(x_m,y_m,z_m), movable part is installed n characteristic point P '₁(x��₁,y��₁,z��₁),P��₂(x��₂,y��₂,z��₂)...P��_n(x��_n,y��_n,z��_n), namely known this m+n point assembled after three-dimensional coordinate on trim designs model.

3rd step: measure the three-dimensional coordinate at optics control point;

Constitute double camera Stereo Vision Measurement System after camera calibration, obtain the location parameter between the distortion parameter of camera and two cameras. Structural member is shot by double camera simultaneously, and left and right camera carries out image procossing and feature extraction, obtains the image coordinate of optical characteristic point, i.e. (x_L,y_L) and (x_L,y_L), resolve all optics control point 3 d space coordinate (X, Y, Z) under camera coordinates system according to equation below.

$\left\{\begin{array}{c}X={\mathrm{Zx}}_L/f_L\\ Y={\mathrm{Zy}}_L/f_L\\ Z=\frac{f_L(f_R{t}_x-x_L{t}_z)}{x_R(r_7{x}_L+r_8{y}_L+r_9{f}_L)-f_R(r_1{x}_L+r_2{y}_L+r_3{f}_L)}\\ =\frac{f_L(f_R{t}_y-y_L{t}_z)}{y_R(r_7{x}_L+r_8{y}_L+r_9{f}_L)-f_R(r_4{x}_L+r_5{y}_L+r_6{f}_L)}\end{array}\right.---\left(1\right)$

In formula: left camera coordinates system O_L-X_LY_LZ_LFor world coordinate system, right camera coordinates system O_R-X_RY_RZ_RRepresent relative to the spatial relationship spin matrix R and translation matrix T of world coordinate system, (x_L,y_L) and (x_L,y_L) picpointed coordinate of respectively left and right image. f_L��f_RRepresent the effective focal length of left and right camera respectively; f_L��f_RRepresent the effective focal length of left and right camera respectively; Matrix R_CameraAnd T_CameraFor:

4th step: set up assembling coordinate system;

Three-dimensional coordinate on designing a model after having been assembled by second step m fixture characteristic point of acquisition and n movable part characteristic point; This m+n some three-dimensional coordinate under camera coordinates system is measured by the 3rd pacing. By m the fixture characteristic point coordinate under Two coordinate system respectively, utilize Coordinate Conversion principle, resolve the transformational relation tried to achieve between assembling coordinate system and camera coordinates system, if transition matrix is R_Coordinate. Utilize transition matrix R_{Sit Mark}Be transformed into by n movable part characteristic point three-dimensional coordinate under camera coordinates system under assembling coordinate system, then now the three-dimensional coordinate of all characteristic points is all unified under assembling coordinate system.

5th step: the attitude misalignment between real-time resolving assembling structural member;

By rigid body translation principle it can be seen that gestures of object change time object on point coordinates meet:

P��_i=R P_i+T(2)

Wherein R is spin matrix, and T is translation matrix, P_iFor the three-dimensional coordinate of characteristic point, P before attitudes vibration_i' for the three-dimensional coordinate of characteristic point after attitudes vibration.

$\begin{array}{c}R=\left[\begin{array}{ccc}{b}_{11}& b_{12}& b_{13}\\ b_{21}& b_{22}& b_{23}\\ b_{31}& b_{32}& b_{33}\end{array}\right]=R_z{R}_y{R}_x\\ =\left[\begin{array}{ccc}\cos \mathrm{\β}\cos \mathrm{\γ}& -\cos \mathrm{\α}\sin \mathrm{\γ}+\sin \mathrm{\α}\sin \mathrm{\β\γ}\cos & \sin \mathrm{\α}\sin \mathrm{\γ}+\cos \mathrm{\α}\sin \mathrm{\β}\cos \mathrm{\γ}\\ \cos \mathrm{\β}\sin \mathrm{\γ}& \cos \mathrm{\α}\cos \mathrm{\γ}+\sin \mathrm{\α}\sin \mathrm{\β}\sin \mathrm{\γ}& -\sin \mathrm{\α}\cos \mathrm{\γ}+\cos \mathrm{\α}\sin \mathrm{\β}\sin \mathrm{\γ}\\ -\sin \mathrm{\β}& \sin \mathrm{\α}\cos \mathrm{\β}& \cos \mathrm{\α}\cos \mathrm{\β}\end{array}\right]\end{array}$

Wherein, ��, beta, gamma is respectively around x, y, the anglec of rotation of z-axis;

$T=\left[\begin{array}{c}{X}_t\\ Y_t\\ Z_t\end{array}\right]$

Wherein X_t, Y_t, Z_tRespectively along x, y, the translational movement of z-axis.

In assembling process, by double camera vision measurement system, the characteristic point on structural member being measured in real time, the movement parts characteristic point recorded in each moment coordinate under assembling coordinate system brings P in formula (2) into_i, bring movement parts characteristic point coordinate under trim designs model in formula (2) P_i', utilize above-mentioned geometrical relationship to set up equation and can solve spin matrix and translation matrix, it is achieved attitude algorithm, export six degree of freedom assembling deviation ��, beta, gamma, X_t, Y_t, Z_t. This deviation value can feed back to control system, instructs the movement executing mechanism of assembling docking to complete assembling process.

The present invention adopts and arranges and measure that camera is gone forward side by side that rower is fixed, installed optics control point on tested structural member, the three-dimensional coordinate of measuring optics control point, the attitude misalignment set up between assembling coordinate system and real-time resolving assembling structural member set step, adopt double camera principle of stereoscopic vision, solve the six-degree-of-freedom posture deviation on-line measurement problem in large-sized structural parts assembling docking operation, the method measurement device is simple, cost is low, and can detect the elastic deformation of large-sized structural parts. Experimental result shows, when alignment measurement space is 5m �� 5m �� 5m, and the feature luminous point of employing reaches 5, the root-mean-square error of this attitude measurement method is maintained within 0.05 ��, and refresh rate reaches 200 frames/second, and data are accurately and reliably.

Claims (8)

1. a large-sized structural parts assembling docking measuring method for view-based access control model principle, comprises the steps:

The first step: arrange that measuring camera goes forward side by side rower calmly;

Second step: install optics control point on tested structural member;

3rd step: measure the three-dimensional coordinate at optics control point;

4th step: set up assembling coordinate system;

5th step: the attitude misalignment between real-time resolving assembling structural member.

2. the large-sized structural parts assembling docking measuring method of a kind of view-based access control model principle according to claim 1, it is characterized in that: described layout is measured camera and carries out demarcating steps, two structural members to be assembled are placed in assembly shop and adjust in unwheeling, defining one of them structural member is fixture, in assembling process, position and attitude does not all change, another structural member is movable part, by adjusting position and the attitude of movable part so that it is alignment fixture completes assembling; Arrange two measurement cameras, it is desirable to two camera intersections are arranged, intersection angle about about 60 ��, and make intersection visual field cover two structural member assembly spaces; After fixed camera, the inner parameter of camera and external parameter are demarcated, so as to constitute double camera to measure system.

3. the large-sized structural parts assembling docking measuring method of a kind of view-based access control model principle according to claim 2, it is characterized in that: described installs optics control point step on tested structural member, two tested structural members to be assembled fix several optics control point respectively, it is simple to vision measurement system carries out feature recognition and extraction; Above-mentioned optics control point dispersed placement, covered structure part profile.

4. the large-sized structural parts assembling docking measuring method of a kind of view-based access control model principle according to claim 1, it is characterized in that: the described three-dimensional coordinate step measuring optics control point, constitute double camera Stereo Vision Measurement System after camera calibration, obtain the location parameter between the distortion parameter of camera and two cameras.

5. the large-sized structural parts assembling docking measuring method of a kind of view-based access control model principle according to claim 4, it is characterized in that: the described three-dimensional coordinate step measuring optics control point, structural member is shot by double camera simultaneously, left and right camera is carried out image procossing and feature extraction, obtain the image coordinate of optical characteristic point, i.e. (x_L,y_L) and (x_L,y_L), resolve all optics control point 3 d space coordinate (X, Y, Z) under camera coordinates system according to equation below;

$\left\{\begin{array}{c}X={\mathrm{Zx}}_L/f_L\\ Y={\mathrm{Zy}}_L/f_L\\ Z=\frac{f_L(f_R{t}_x-x_L{t}_z)}{x_R(r_7{x}_L+r_8{y}_L+r_9{f}_L)-f_R(r_1{x}_L+r_2{y}_L+r_3{f}_L)}\\ =\frac{f_L(f_R{t}_y-y_L{t}_z)}{y_R(r_7{x}_L+r_8{y}_L+r_9{f}_L)-f_R(r_4{x}_L+r_5{y}_L+r_6{y}_L)}\end{array}\right.---\left(1\right)$

In formula: left camera coordinates system O_L-X_LY_LZ_LFor world coordinate system, right camera coordinates system O_R-X_RY_RZ_RRepresent relative to the spatial relationship spin matrix R and translation matrix T of world coordinate system, (x_L,y_L) and (x_L,y_L) picpointed coordinate of respectively left and right image; f_L��f_RRepresent the effective focal length of left and right camera respectively; f_L��f_RRepresent the effective focal length of left and right camera respectively; Matrix R_CameraAnd T_CameraFor:

6. the large-sized structural parts assembling docking measuring method of a kind of view-based access control model principle according to claim 1, it is characterized in that: described foundation assembling coordinate system step, second step m fixture characteristic point of acquisition and n movable part the characteristic point three-dimensional coordinate on designing a model after having assembled; This m+n some three-dimensional coordinate under camera coordinates system is measured by the 3rd pacing; By m the fixture characteristic point coordinate under Two coordinate system respectively, utilize Coordinate Conversion principle, resolve the transformational relation tried to achieve between assembling coordinate system and camera coordinates system, if transition matrix is R_Coordinate; Utilize transition matrix R_CoordinateBe transformed into by n movable part characteristic point three-dimensional coordinate under camera coordinates system under assembling coordinate system, then now the three-dimensional coordinate of all characteristic points is all unified under assembling coordinate system.

7. the large-sized structural parts assembling docking measuring method of a kind of view-based access control model principle according to claim 6, it is characterized in that: the described attitude misalignment step between real-time resolving assembling structural member, in assembling process, by double camera vision measurement system, the characteristic point on structural member being measured in real time, the movement parts characteristic point recorded in each moment coordinate under assembling coordinate system is as P_i, movement parts characteristic point coordinate under trim designs model is as P_i', use equation below to calculate spin matrix R and translation matrix T:

P_i'=R P_i+T(2)

Utilize spin matrix and translation matrix, it is achieved attitude algorithm, export six degree of freedom assembling deviation ��, beta, gamma, X_t, Y_t��

8. the large-sized structural parts assembling docking measuring method of a kind of view-based access control model principle according to claim 7, it is characterised in that: the described attitude misalignment step between real-time resolving assembling structural member,

$R=\left[\begin{array}{ccc}\cos \mathrm{\β}\cos \mathrm{\γ}& -\cos \mathrm{\α}\sin \mathrm{\γ}+\sin \mathrm{\α}\sin \mathrm{\β}\cos \mathrm{\γ}& \sin \mathrm{\α}\sin \mathrm{\γ}+\cos \mathrm{\α}\sin \mathrm{\β}\cos \mathrm{\γ}\\ \cos \mathrm{\β}\sin \mathrm{\γ}& \cos \mathrm{\α}\cos \mathrm{\γ}+\sin \mathrm{\α}\sin \mathrm{\β}\sin \mathrm{\γ}& -\sin \mathrm{\α}\cos \mathrm{\γ}+\cos \mathrm{\α}\sin \mathrm{\β}\sin \mathrm{\γ}\\ -\sin \mathrm{\β}& \sin \mathrm{\α}\cos \mathrm{\β}& \cos \mathrm{\α}\cos \mathrm{\β}\end{array}\right];$

Wherein ��, beta, gamma is respectively around x, y, the anglec of rotation of z-axis;

$T=\left[\begin{array}{c}{X}_t\\ Y_t\\ Z_t\end{array}\right];$

Wherein X_t, Y_t, Z_tRespectively along x, y, the translational movement of z-axis.