CN106971408A - A kind of camera marking method based on space-time conversion thought - Google Patents

Abstract

A kind of camera marking method based on space-time conversion thought of the present invention belongs to Computer Vision Detection and field of image detection, is related to a kind of camera marking method based on space-time conversion thought.The three-dimensional rotation table device of laser constitution that this method passes through two turntables being mutually located at a certain angle and a turntable installed therein, pass through the separate motion control laser of two turntables, laser spots are made to be scanned on wall, the conversion idea in passage time and space again, obtain the distance scanned and the relational expression of turntable rotational time, laser dot image is shot at correspondence time point using video camera, obtain a large amount of corresponding two-dimensional space characteristic points, then according to national forest park in Xiaokeng principle, the intrinsic parameter and outer parameter of video camera are drawn respectively.This method need not take excessive space when performing demarcation, the speed of projection target spot also quickly, saves the nominal time, therefore can effectively adapt to processing site environment.

Description

A kind of camera marking method based on space-time conversion thought

Technical field

The invention belongs to Computer Vision Detection and field of image detection, it is related to a kind of taking the photograph based on space-time conversion thought Camera calibration method.

Background technology

Computer vision and image detection application in, for determine space object surface point three-dimensional geometry position and its Correlation between two-dimentional corresponding points in the picture, it is necessary to set up the geometrical model of video camera imaging, these geometrical models ginseng Number is exactly camera parameters.These parameters must can just be obtained by experiment with calculating in most conditions, and this solves ginseng Several processes is camera calibration.Therefore the demarcation of camera parameters is unusual the key link in vision-based detection, and it is demarcated As a result precision and the stability of algorithm directly affects the accuracy that camera operation produces result.Therefore, video camera mark is simplified It is fixed, improve the emphasis place that stated accuracy is current research work.But existing conventional demarcation mode is almost required for demarcation Plate or other three-dimensional objects of reference assist to demarcate as target, are then shot by the way that video camera is multigroup, to obtain the correlation of video camera Parameter.But when carrying out in-site measurement, due to complex environment, the time of demarcation and space are extremely limited, the shifting of measured object It is dynamic that conventional scaling method can be caused to be difficult to fast and effectively complete demarcation, cause measurement period to extend, the real-time drop of measurement Low and vision measurement precise decreasing.

Conventional scaling method is that Zhang Zhengyou exists at present《A Flexible New Technique for Camera Calibration》The scaling method based on plane reference plate proposed in one text, this method is between tradition demarcation and self-calibration Between a kind of method, it only needs to video camera and shoots several pictures from different directions to some scaling board, by scaling board Corresponding relation between the picture point of each characteristic point and its image plane, i.e., often the homography matrix of piece image carries out the mark of video camera It is fixed, there is relatively broad application.This method has degree of precision, but needs scaling board to carry out auxiliary calibration, and stated accuracy is relied on In the precision of scaling board, and scaling board is expensive, while Zhang Shi standardizations algorithm is complicated, time-consuming for demarcation, is needed during demarcation Substantial amounts of space is occupied, it is inconvenient when measuring at the scene, therefore be not suitable for the live on-line tuning of camera parameters.

The content of the invention

The technical problems to be solved by the invention are to overcome the shortcomings of existing scaling method, for lacking at commercial measurement scene The situation of few effective camera marking method, invents a kind of camera marking method based on space-time conversion thought.This method Using laser scanning in wall plane, a large amount of position feature point information are obtained, according to the characteristic point of acquisition to video camera Parameter is demarcated, and acquisition includes principal point coordinate and equivalent focal length of the optical axis by image plane, the spin matrix of video camera with And the parameter such as the translation matrix between video camera.This method need not take excessive space when performing demarcation, project the speed of target spot Also quickly, the nominal time is saved, therefore processing site environment can be effectively adapted to.

The present invention adopts the technical scheme that a kind of camera marking method based on space-time conversion thought, and this method passes through The three-dimensional rotation table device of turntable and the laser constitution of a turntable installed therein that two are mutually located at a certain angle, passes through The separate motion control laser of two turntables, makes laser spots be scanned on wall, then passage time and space conversion Thought, obtains the distance scanned and the relational expression of turntable rotational time, using video camera at correspondence time point to laser dot image Shot, obtain a large amount of corresponding two-dimensional space characteristic points, then according to national forest park in Xiaokeng principle, video camera is drawn respectively Intrinsic parameter and outer parameter, method comprises the following steps that：

Step 1：The assembling of three-dimensional rotation table device

In three-dimensional rotation table device, laser 3 is fixed therein by the mutual location and installation at a certain angle of upper and lower turntable 1,2 On one turntable, by the separate motion control laser 3 of two turntables, realize laser 3 in any of the plane of wall 6 Place's projection laser spots, while being shot using video camera 4 to laser target dot image, are obtained and include laser target point feature point information Image；

Step 2：Characteristic point on wall is repeatedly shot using video camera timesharing, multiple series of images information data is obtained.

Multigroup laser target spot is projected in target wall using three-dimensional turret plant, while repeatedly shooting, in the every of shooting In one group of image, the actual size of characteristic point in Each point in time shooting image is obtained by space-time conversion relational expression.

Formula (1) is the space-time conversion relational expression being derived from according to accompanying drawing 2, wherein：L is characterized reality a little with initial point Border size, as can be seen that the unknown quantity is the imaginary circle being made up of rotary laser target spot, the speed on wall from accompanying drawing 2 What the integration of projection was obtained, its velocity magnitude is w r, and projected angle is α-θ, it is therefore desirable to record measure other correlated variables r, θ, L、α、w、t；Wherein r is the distance that initial point arrives laser, and θ is initial spot speed and L angle, and α is characterized spot speed and L Angle, w is the angular velocity of rotation of turntable, and t is the rotational time of turntable.Brought into formula (1), can solved using known parameters The relation of characteristic point and the actual size of initial point in calibration process, i.e., specific time point and space characteristics point position；According to Multigroup projection dot image that camera is not shot in the same time, can solve the position letter of the characteristic point in each group of camera calibration Breath.

Any characteristic point, using wall plane as X O Y planes, sets up space multistory coordinate as the origin of coordinates using on wall System, referred to as world coordinate system；Due to each characteristic point actual size on wall, it is known that so each characteristic point is in the world on wall Known to X-axis coordinate and Y-axis coordinate under coordinate system.When metope is approximately ideal plane, Z axis coordinate is zero；When metope not During ideal plane, the characteristic point that laser scans acquisition has three-dimensional information, and each characteristic point Z-direction information can be sat by pole Mark (L, β) and represent that the solution of parameter 3 can obtain equation below with reference to the accompanying drawings wherein in polar coordinates：

β=90 °-α-arccos ((L²+r²-r₁ ²)/2rL) (3)

As shown in formula (2), the point on non-ideal plane wall is assumed to be arbitrary characteristics point, its polar coordinates is (L, β), As shown in Figure 3, first with the length r and r of measurement₁The cosine law, obtain the polar diameter L of note coordinate, then pass through characteristic quantity The geometrical relationship of triangle, obtains polar coordinates polar angle β；Finally according to projection of the polar coordinates on X O Y planes of characteristic point, obtain The X-axis coordinate and Y-axis coordinate of characteristic point are obtained, so as to obtain the geometric properties information of characteristic point.

Step 3：Set up Optimized model optimization calibrating parameters

Camera calibration is using classical national forest park in Xiaokeng, by step 2 at the geometric properties information of each group of acquisition Bring simultaneous in following pin-point model after reason into and obtain multigroup equation, you can solve intrinsic parameter, the camera coordinate system of video camera With the spin matrix and translation vector of world coordinate system, that is, the related intrinsic parameter of camera and the outer parameter at demarcation scene are obtained, The expression formula of pin-point model is as follows

Wherein, (X_w,Y_w,Z_w, 1)^TThe homogeneous coordinates for being spatial point in world coordinate system, (u, v, 1)^TFor corresponding image Picture point pixel coordinate system ο₀Homogeneous coordinates in uv, α_x=f/dx is ο₀Scale factor in uv coordinate systems on u axles, α_y=f/dy For coordinate system ο₀Scale factor in uv on v axles, f is camera lens focal length, d_xWith d_yRespectively horizontal, the vertical physics chi of pixel It is very little, (u₀,v₀) it is main point coordinates, ρ_cFor proportionality coefficient, K is intrinsic parameters of the camera matrix, and [R | t] is the outside ginseng of video camera Matrix number, wherein, R is spin matrix, and t is translation vector.

Afterwards, using video camera intrinsic parameter, camera coordinate system and world coordinate system spin matrix and translation vector Solve all characteristic point re-projection coordinates in target wallSpecific formula is as follows：

Wherein, r_ijFor the element on spin matrix R the i-th row, jth row, translation vector t=(t₁,t₂,t₃)^T, f_xTo take the photograph Camera horizontal scaling factor, f_yFor video camera vertical scaling factor, ρ₀For abscissa of the principal point under pixel coordinate system, λ₀Based on Ordinate of the point under pixel coordinate system, (X_W,Y_W,Z_W) it is characterized the coordinate a little under world coordinate system.

According to known distortion factor, the re-projection picpointed coordinate that actual photographed is obtainedIt is corrected to corresponding ideal Picpointed coordinate (u_n,v_n)；Deviation of the Optimized model by iteration minimization re-projection picpointed coordinate and ideal image point coordinate is set up, Objective optimization function is：

Using LM nonlinear optimization algorithms, it is changed into symmetric positive definite matrix by Hessian gusts and solves, it is corresponding when deviation is minimum Parameter is the computer vision system camera parameters after optimization.

Mark need not can be ensured using the high-precision equipment such as auxiliary calibration plate the beneficial effects of the invention are as follows scaling method Determine precision, space-consuming is not needed when performing demarcation, the speed of projection target spot also quickly, saves the nominal time.It therefore, it can have Effect ground adapts to processing site environment, is effectively improved the effect of field calibration, improves demarcation efficiency.

Brief description of the drawings

Fig. 1 for the camera marking method of space-time conversion thought schematic device, wherein, turntable under the upper turntables of 1-, 2-, 3- lasers, 4- video cameras, 5- supports, 6- walls.

Fig. 2 is space-time conversion inference schematic diagram.Wherein, L is characterized actual size a little with initial point, and r arrives for initial point The distance of laser, θ is initial spot speed and L angle, and α is initial characteristicses spot speed and L angle, and w is the rotation of turntable Angular speed, t is the rotational time of turntable.

Fig. 3 is non-ideal plane wall space-time conversion inference schematic diagram.Wherein, L is characterized actual chi a little with initial point Very little, r is the distance that initial point arrives laser, and point P takes up an official post meaning point, r for non-ideal plane wall₁For arbitrfary point to laser away from From θ is initial spot speed and L angle, and β is arbitrfary point and the angle of preferable metope plane, and α is initial characteristicses spot speed and L Angle, w be turntable angular velocity of rotation, t be turntable rotational time.

Embodiment

Describe the implementation of the present invention in detail with technical scheme below in conjunction with the accompanying drawings.

Fig. 1 is the schematic device of camera marking method, and method is comprised the following steps that：

Step 1：Three-dimensional turret plant is installed

As shown in Figure 1, three-dimensional turret plant is mutually located at a certain angle by upper turntable 1 and lower turntable 2, will when using Laser 3 is fixed on upper turntable 1, by the respective motion of two turntables, realizes that laser 3 is thrown in any place of the plane of wall 6 Penetrate the purpose of laser spots.

Step 2：The characteristic point on wall 6 is repeatedly shot using the timesharing of video camera 4, multiple series of images information data is obtained.It is right The image of shooting carries out data processing, in each group of image of shooting, on business formula (1) be with reference to the accompanying drawings 2 be derived from when Empty conversion relational expression, the formula can solve the position relationship of calibration process intermediate station rotational time and space characteristics point, not Camera shoots multigroup projection dot image in the same time, under the conditions of known to above-mentioned each parameter, can solve each group of video camera mark The positional information of characteristic point in fixed.

Any characteristic point is the origin of coordinates using on wall 6, using wall plane as XOY plane, sets up space multistory coordinate system, Referred to as world coordinate system；Due to each characteristic point actual size on wall, it is known that so each characteristic point is sat in the world on wall 6 Known to X-axis coordinate and Y-axis coordinate under mark system.When metope is approximately ideal plane, Z axis coordinate is zero.When metope is not reason When thinking plane, equally X-axis coordinate and Y-axis of each characteristic point under world coordinate system on wall can be derived with reference to the accompanying drawings 2 Coordinate, and each characteristic point Z axis polar coordinates.If metope is not ideal plane, XOY plane can not be obtained according to above-mentioned formula Relevant parameter, it is necessary to first obtain the polar coordinates (L, β) of each characteristic point Z axis, solved using formula (2), (3), according to each The polar coordinates of characteristic point, are projected on XOY plane, are obtained the X-axis coordinate and Y-axis coordinate of characteristic point, are obtained characteristic point All characteristic informations.

Step 3：Set up Optimized model optimization calibrating parameters

The classical national forest park in Xiaokeng of camera calibration use, the model expression such as formula (4), wherein, (X_w,Y_w,Z_w, 1)^TThe homogeneous coordinates for being spatial point in world coordinate system, (u, v, 1)^TFor corresponding image picture point pixel coordinate system ο₀In uv Homogeneous coordinates, obtain respectively obtaining multigroup parameter in characteristic information from step 2 after video camera shooting image and image procossing (X_w,Y_w,Z_w, 1)^T(u, v, 1)^T, then by constituting sufficient amount of equation with pin-point model formula (3) simultaneous, can solve The inside and outside parameter of camera is obtained, including：α_x=f/dx is ο₀Scale factor in uv coordinate systems on u axles, α_y=f/dy is coordinate system ο₀Scale factor in uv on v axles, f is camera lens focal length, d_xWith d_yRespectively horizontal, the vertical physical size of pixel, (u₀,v₀) For main point coordinates, ρ_cFor proportionality coefficient, K is intrinsic parameters of the camera matrix, and [R | t] is the external parameter matrix of video camera, its In, R is spin matrix, and t is translation vector.

By the inside and outside parameter of the video camera obtained before, substitute into formula (4) and obtain all features in solution target wall Point re-projection actual coordinateThe re-projection picpointed coordinate obtained due to actual photographedIn the presence of certain distortion, need Ideal image point coordinate (u is corrected to according to known distortion parameter_n,v_n)；Then Optimized model is set up, passes through iteration pole The deviation of smallization re-projection picpointed coordinate and ideal image point coordinate, its objective optimization function is：

Using LM nonlinear optimization algorithms, it is changed into symmetric positive definite matrix by Hessian gusts and solves, obtained when deviation is minimum Correspondence inside and outside parameter, the computer vision system camera parameters as after final optimization pass.

Claims (1)

1. a kind of camera marking method based on space-time conversion thought, it is characterized in that, this method by two at a certain angle The three-dimensional rotation table device of turntable and the laser constitution of a turntable installed therein being mutually located, it is mutually only by two turntables Vertical motion control laser, make laser spots be scanned on wall, then passage time and space conversion idea, obtain and scan The relational expression of distance and turntable rotational time, is shot using video camera at correspondence time point to laser dot image, obtains big Corresponding two-dimensional space characteristic point is measured, then according to national forest park in Xiaokeng principle, the intrinsic parameter and outer ginseng of video camera are drawn respectively Number, method is comprised the following steps that：

Step 1：The assembling of three-dimensional rotation table device

In three-dimensional rotation table device, laser (3) is fixed therein by the mutual location and installation at a certain angle of upper and lower turntable (1,2) On one turntable, by the separate motion control laser (3) of two turntables, realize laser (3) in wall (6) plane Any place projection laser spots, while shot using video camera (4) to laser target dot image, obtain comprising laser target spot spy Levy the image of an information；

Step 2：Characteristic point on wall is repeatedly shot using video camera (4) timesharing, multiple series of images information data is obtained；In laser spots Repeatedly shot while projection, in each group of image of shooting, Each point in time is obtained by space-time conversion relational expression and shot Characteristic point actual size in image；Solve the position relationship of calibration process intermediate station rotational time and space characteristics point, when dally Changing relational expression is：

$L=\underset{0}{\overset{t}{\∫}}wrcos(\mathrm{\α}-\mathrm{\θ})dt---\left(1\right)$

Wherein：L is characterized actual size a little with initial point, and r is distance of the initial point to laser, and θ is initial spot speed and L Angle, α is characterized spot speed and L angle, and w is the angular velocity of rotation of turntable, and t is the rotational time of turntable；

Video camera does not shoot multigroup projection dot image in the same time, under the conditions of known to above-mentioned each parameter, each group of solution is taken the photograph The positional information of characteristic point in camera calibration；

Using any characteristic point on wall (6) as the origin of coordinates, using wall (6) plane as XOY plane, space multistory coordinate is set up System, referred to as world coordinate system；Due to each characteristic point actual size on wall, it is known that so each characteristic point is in the world on wall Known to X-axis coordinate and Y-axis coordinate under coordinate system；When metope is approximately ideal plane, Z axis coordinate is zero；When metope is not During ideal plane, the characteristic point that laser is scanned has three-dimensional information, and each characteristic point Z axis is represented by polar coordinates (L, θ), wherein The solution formula of parameter in polar coordinates：

$L=\sqrt{r^2+r_1^2-2r.r_{1}cos\left(wt\right)}---\left(2\right)$

β=90 °-α-arccos ((L²+r²-r₁ ²)/2rL) (3)

According to the polar coordinates of each characteristic point, projected on XOY plane, obtain the X-axis coordinate and Y-axis coordinate of characteristic point, Obtain all characteristic informations of characteristic point；

Step 3：Set up Optimized model optimization calibrating parameters

Camera calibration is using classical national forest park in Xiaokeng, and the expression formula of the model is as follows：

Wherein, (X_w,Y_w,Z_w, 1)^TThe homogeneous coordinates for being spatial point in world coordinate system, (u, v, 1)^TFor corresponding image picture point Pixel coordinate system o₀Homogeneous coordinates in uv, α_x=f/dx is o₀Scale factor in uv coordinate systems on u axles, α_y=f/dy is seat Mark system o₀Scale factor in uv on v axles, f is camera lens focal length, d_xWith d_yRespectively horizontal, the vertical physical size of pixel, (u₀,v₀) it is main point coordinates, ρ_cFor proportionality coefficient；If K is intrinsic parameters of the camera matrix, [R | t] is the external parameter of video camera Matrix, wherein, R is spin matrix, and t is translation vector；Intrinsic parameters of the camera includes principal point coordinate (u₀,v₀), scale factor α_x、α_y, coefficient of radial distortion k₁、k₂With tangential distortion coefficient p₁、p₂；Video camera external parameter is camera coordinate system relative to generation The orientation of boundary's coordinate system, including spin matrix R and translation vector T；

Scaling board is solved using the spin matrix and translation vector of the intrinsic parameter, camera coordinate system and world coordinate system of video camera Upper all characteristic point re-projection coordinatesSpecific algorithm is as follows：

$\left\{\begin{array}{c}\hat{u}=\frac{\left(f_x{r}_{11}+u_0{r}_{31}\right)X_W+\left(f_x{r}_{12}+u_0{r}_{32}\right)Y_W+\left(f_x{r}_{13}+u_0{r}_{33}\right)Z_W+f_x{t}_1+u_0{t}_3}{r_{31}{X}_W+r_{32}{Y}_W+r_{33}{Z}_W+t_3}\\ \hat{v}=\frac{\left(f_y{r}_{21}+v_0{r}_{31}\right)X_W+\left(f_y{r}_{22}+v_0{r}_{32}\right)Y_W+\left(f_y{r}_{23}+v_0{r}_{33}\right)Z_W+f_y{t}_2+v_0{t}_3}{r_{31}{X}_W+r_{32}{Y}_W+r_{33}{Z}_W+t_3}\end{array}\right.---\left(5\right)$

Wherein, r_ijFor the element on spin matrix R the i-th row, jth row, translation vector t=(t₁,t₂,t₃)^T, f_xIt is horizontal for video camera To scale factor, f_yFor video camera vertical scaling factor, ρ₀For abscissa of the principal point under pixel coordinate system, λ₀It is principal point in picture Ordinate under plain coordinate system, (X_W,Y_W,Z_W) it is characterized the coordinate a little under world coordinate system；

According to known distortion factor, the picpointed coordinate (u ' that actual photographed is obtained_n,v′_n) it is corrected to corresponding ideal image point coordinate (u_n,v_n)；Set up deviation of the Optimized model by iteration minimization re-projection picpointed coordinate and ideal image point coordinate, objective optimization Function is：

$\min \left(\underset{n=1}{\overset{m}{\mathrm{\Σ}}}\right({\left(u_n-{\hat{u}}_n\right)}^2+{\left(v_n-{\hat{v}}_n\right)}^2\left)\right)---\left(6\right)$

Using LM nonlinear optimization algorithms, it is changed into symmetric positive definite matrix by Hessian gusts and solves, the corresponding parameter when deviation is minimum Computer vision system camera parameters after as optimizing.