阅读说明：本技术 基于阵列式相机的高空间分辨率三维数字图像测量方法 (Array camera-based high-spatial-resolution three-dimensional digital image measuring method ) 是由 邵新星 魏康 何小元 于 2019-09-24 设计创作，主要内容包括：本发明提供了一种基于阵列式相机的高空间分辨率三维数字图像相关测量方法,该方法包括如下步骤：(1)加工并进行N×N(N≧2)二维相机阵列的安装,完成对于左右两组二维相机阵列的布置方案；(2)在被测试样表面制备随机散斑图案；(3)搭建测量系统,组成三维相机阵列；(4)使用编码标定板对三维相机阵列中的相机参数进行标定；(5)通过同步采集装置使得三维相机阵列中相机同时采集试样表面图像,并获取所述试样表面图像的位置信息载体；(6)对三维相机阵列采集的试样表面图像的位置信息载体进行阵列相机图像之间的时序匹配和立体匹配；(7)实现位移场数据的校正和融合拼接,并在此基础上求解获取应变计算。本发明的方法具有高空间分辨率、操作简便的特点。(The invention provides a high-spatial-resolution three-dimensional digital image correlation measurement method based on an array camera, which comprises the following steps of: (1) processing and installing the N multiplied by N (N is larger than or equal to 2) two-dimensional camera arrays to finish the arrangement scheme of the left and right two-dimensional camera arrays; (2) preparing a random speckle pattern on the surface of a sample to be tested; (3) building a measuring system to form a three-dimensional camera array; (4) calibrating camera parameters in the three-dimensional camera array by using a coding calibration plate; (5) enabling cameras in the three-dimensional camera array to simultaneously acquire surface images of the sample through a synchronous acquisition device, and acquiring a position information carrier of the surface images of the sample; (6) carrying out time sequence matching and stereo matching between array camera images on a position information carrier of a sample surface image acquired by a three-dimensional camera array; (7) and correcting and fusion splicing of displacement field data are realized, and strain calculation is obtained by solving on the basis. The method has the characteristics of high spatial resolution and simple and convenient operation.)

1. A high spatial resolution three-dimensional digital image correlation measurement method based on an array camera is characterized by comprising the following steps:

(1) processing and installing two groups of NxN (N is larger than or equal to 2) two-dimensional camera arrays, wherein the two groups of two-dimensional camera arrays form a three-dimensional camera array, and finishing the arrangement scheme of the left and right two groups of two-dimensional camera arrays so that the measured area can be continuously imaged in the view fields of the two groups of two-dimensional camera arrays in the measuring process;

(2) building a three-dimensional camera array digital image correlation measurement system consisting of two groups of two-dimensional camera arrays and a lighting device;

(3) preparing random speckle patterns on the surface of a sample to be tested, and using the prepared speckle coding patterns as position information carriers;

(4) calibrating internal parameters and external parameters of a camera in the three-dimensional camera array by using a coding calibration plate, wherein the internal parameter calibration parameters comprise: principal point location (u)₀,v₀) Lens focal length f, pixel size (S)_x,S_y) And lens distortion parameter matrix D_iThe principal point position is the pixel coordinate of the intersection point of the optical axis of the lens and the target surface, and the lens distortion parameter matrix comprises a 6-order radial distortion parameter K₁、K₂、K₃、K₄、K₅、K₆And 2 tangential distortion number P₁、P₂(ii) a The external reference calibration parameters comprise: the relative external parameters of the cameras in the two-dimensional camera arrays and the relative external parameters of the cameras between the two-dimensional camera arrays, each external parameter of the cameras comprises 6 parameters n_x、n_y、n_z、t_x、t_y、t_z；

(5) Adjusting the observation position and the observation angle of the three-dimensional camera array, ensuring that the detected area is completely imaged in the three-dimensional camera array, enabling the cameras in the three-dimensional camera array to simultaneously acquire the surface image of the sample through the synchronous acquisition device, and acquiring a position information carrier of the surface image of the sample;

(6) based on a digital image correlation technology, carrying out time sequence matching and stereo matching between array camera images on a position information carrier of the sample surface image acquired by the three-dimensional camera array in the step (5);

(7) performing three-dimensional reconstruction on the calculation points in the array image according to the camera internal parameters and the array camera external parameters obtained by calculation in the step (4) and the matching result in the step (6) to obtain point cloud data, and performing calculation on the point cloud data before and after deformation to obtain three-dimensional displacement field data;

(8) and (4) calculating to obtain external parameters of the phase in the array according to the step (4) and point cloud data obtained by three-dimensional reconstruction in the step (7), realizing correction and fusion splicing of displacement field data, and solving to obtain full field strain on the basis.

2. The array-type camera-based high spatial resolution three-dimensional digital image measuring method according to claim 1, wherein in the step (1), during the arrangement of the two-dimensional camera arrays, when an object is displaced from the plane by Δ Z, for two cameras in the array with a working distance Z, the following relationship is obtained according to the imaging geometry:

wherein x and y are image coordinates, Δ x is ImageSize-OverlapSize, ImageSize is the transverse size of the image, OverlapSize is the transverse size of the overlapping area, and subscripts of 0 and 1 respectively represent two adjacent cameras in the array, and should be used for ensuring data continuity

3. The array camera-based high spatial resolution three-dimensional digital image measuring method according to claim 1, wherein in the step (4) of calibrating by using the code calibration plate, both the marking point code and the digital speckle code can be used for calibration; the internal parameters of the cameras in the three-dimensional camera array can be calibrated by directly using a plane calibration method; the external parameters of the cameras in the three-dimensional camera array need to calibrate the relation between a camera coordinate system and a calibration board world coordinate system, and then the external parameters of the cameras in the array and the external parameters of the cameras among the array are solved according to the transformation of the coordinate system:

establishing a world coordinate system on a coding calibration board, wherein an external reference relation exists between the camera coordinate system of each camera and the world coordinate system, and the external reference relation comprises 6 parameters n_x、n_y、n_z、t_x、t_y、t_zAnd transforming the rotation vector therein by Rodrigues to obtain a rotation matrix:

according to the external reference relationship between one camera coordinate system C1 and the world coordinate system W and the external reference relationship between the other camera coordinate system C2 and the world coordinate system W, the external reference relationship between the two camera coordinate systems C1 and C2 in the three-dimensional camera array can be obtained:

4. the array camera-based high spatial resolution three-dimensional digital image measuring method according to claim 1, wherein in the time sequence matching and stereo matching process in step (6), based on a two-dimensional digital image correlation technique, digital image correlation matching is performed on sub-regions divided in left and right images according to a polar line and deformation constraint matching principle, a matching algorithm in the digital image correlation matching is a reverse combination gauss-newton algorithm, a first-order shape function is selected as a shape function of the digital image correlation time sequence matching, and a second-order shape function is selected as a shape function of the digital image correlation stereo matching, and a parallel computing method is adopted for computing.

5. The array camera-based high spatial resolution three-dimensional digital image measuring method of claim 1, wherein the strain calculation in step (8) is performed by: firstly, carrying out plane fitting on local point cloud data, determining a local three-dimensional coordinate system according to the normal direction of a fitting plane and the projection direction of a transverse axis on the plane, obtaining the local three-dimensional coordinate system, transforming the local displacement field data to the local three-dimensional coordinate system, and solving local strain according to a geometric relationship.

Technical Field

The invention relates to the field of non-contact three-dimensional deformation measurement, in particular to a high-spatial-resolution three-dimensional digital image measuring method based on an array type camera.

Background

The three-dimensional digital image correlation method is a non-contact optical full-field measurement method. The method combines a digital image correlation method and a computer vision principle, and can accurately measure the three-dimensional appearance and deformation of the surface of an object. Compared with other measuring methods based on the interference principle in optical measurement mechanics, the method has the advantages of simple equipment, no need of shock insulation, large measuring range, controllable measuring range and the like. At present, the three-dimensional digital image correlation technology is widely applied to the fields of science and engineering.

For the three-dimensional digital image correlation method, the measurement resolution is mainly limited by the imaging resolution. To improve the imaging resolution, the most straightforward approach is to reduce the pixel size or increase the chip size. However, as the pixel size decreases, the amount of available incoming light also decreases, and image noise increases, severely affecting the measurement results. Meanwhile, the larger chip size causes the image transfer frame rate to be slow, resulting in a decrease in measurement time resolution. High cost image sensors and optics are also a concern. It is not easy to increase the imaging resolution directly from the camera hardware perspective. The three-dimensional camera array can greatly improve the spatial resolution of the deformation measurement of the three-dimensional digital image correlation method and further widen the application range of the camera array measurement technology.

Disclosure of Invention

The purpose of the invention is as follows: the invention aims to provide a high-spatial-resolution three-dimensional digital image correlation measurement method based on an array camera, which greatly improves the spatial resolution of three-dimensional digital image correlation measurement, has the characteristics of high spatial resolution and simple and convenient operation, and has good application prospect in large-view-field and high-precision application scenes.

The above purpose is realized by the following technical scheme:

a high spatial resolution three-dimensional digital image correlation measurement method based on an array camera comprises the following steps:

(1) processing and installing two groups of NxN (N is larger than or equal to 2) two-dimensional camera arrays, wherein the two groups of two-dimensional camera arrays form a three-dimensional camera array, and finishing the arrangement scheme of the left and right two groups of two-dimensional camera arrays so that the measured area can be continuously imaged in the view fields of the two groups of two-dimensional camera arrays in the measuring process;

(2) building a three-dimensional camera array digital image correlation measurement system consisting of two groups of two-dimensional camera arrays and a lighting device;

(3) preparing random speckle patterns on the surface of a sample to be tested, and using the prepared speckle coding patterns as position information carriers;

(4) and calibrating the internal reference and the external reference of the cameras in the three-dimensional camera array by using the code calibration plate. Wherein, the internal reference calibration parameters comprise: principal point location (u)₀,v₀) Lens focal length f, pixel size (S)_x,S_y) And lens distortion parameter matrix D_iWherein the principal point position is the pixel coordinate of the intersection point of the optical axis of the lens and the target surface, and the lens distortion parameter matrix generally comprises 6-order radial distortion parameters K₁、K₂、K₃、K₄、K₅、K₆And 2 tangential distortion number P₁、P₂. Wherein, the external reference calibration parameters comprise: the relative external parameters of the cameras in the two-dimensional camera arrays and the camera external parameters of the cameras between the two-dimensional camera arrays, each camera external parameter comprises 6 parameters n_x、n_y、n_z、t_x、t_y、t_z

(5) Adjusting the observation position and the observation angle of the three-dimensional camera array, ensuring that the detected area is completely imaged in the three-dimensional camera array, enabling the cameras in the three-dimensional camera array to simultaneously acquire the surface image of the sample through the synchronous acquisition device, and acquiring a position information carrier of the surface image of the sample;

(6) based on a digital image correlation technology, carrying out time sequence matching and three-dimensional matching between the array camera images on the position information carriers of the sample surface images acquired by the two groups of camera arrays in the step (5);

(7) performing three-dimensional reconstruction on the calculation points in the array image according to the camera internal parameters and the array camera external parameters obtained by calculation in the step (4) and the matching result in the step (6) to obtain point cloud data, and performing calculation on the point cloud data before and after deformation to obtain three-dimensional displacement field data;

(8) and (4) calculating to obtain external parameters of the phase in the array according to the step (4) and point cloud data obtained by three-dimensional reconstruction in the step (7), realizing correction and fusion splicing of displacement field data, and solving to obtain full field strain on the basis.

Further, in the method of the present invention, in the arrangement process of the two sets of three-dimensional camera arrays in step (1), for an array with a working distance Z, when an object undergoes an out-of-plane displacement Δ Z, the following relation can be obtained according to the imaging geometric relationship:

wherein x and y are image coordinates, Δ x is ImageSize-OverlapSize, ImageSize is the transverse size of the image, OverlapSize is the transverse size of the overlapping area, and subscripts of 0 and 1 respectively represent two adjacent cameras in the array, and should be used for ensuring data continuity

Smaller than the initial OverlapSize, i.e., the initial overlap region is set according to the estimated maximum out-of-plane displacement.

Furthermore, in the method of the present invention, in the calibration process using the code calibration board in step (4), both the marking point code and the digital speckle code can be used for calibration. The internal parameters of the cameras in the three-dimensional camera array can be directly calibrated by using a plane calibration method. The external parameters of the cameras in the three-dimensional camera array need to calibrate the relation between a camera coordinate system and a calibration board world coordinate system, and then the external parameters of the cameras in the array and the external parameters of the cameras among the array are solved according to the transformation of the coordinate system:

establishing a world coordinate system on a coding calibration board, wherein an external reference relation exists between the camera coordinate system of each camera and the world coordinate system, and the external reference relation comprises 6 parameters n_x、n_y、n_z、t_x、t_y、t_zThe rotation vector is transformed by Rodrigues to obtain a rotation matrix

According to the external reference relationship between one camera coordinate system C1 and the world coordinate system W and the external reference relationship between the other camera coordinate system C2 and the world coordinate system W, the external reference relationship between the two camera coordinate systems C1 and C2 in the three-dimensional camera array can be obtained:

further, in the method of the present invention, in the time sequence matching and stereo matching process described in step (6), based on two-dimensional digital image correlation technique, digital image correlation matching is performed on the sub-regions divided in the left and right images according to the polar line and deformation constraint matching principle, the matching algorithm in the digital image correlation matching selects an inverse combination gauss-newton algorithm, the shape function of the digital image correlation time sequence matching selects a first order shape function, and the shape function of the digital image correlation stereo matching selects a second order shape function, and the shape function is calculated by using a parallel calculation method.

Further, in the method of the present invention, the strain calculation in step (8) is performed by: firstly, carrying out plane fitting on local point cloud data, determining a local three-dimensional coordinate system according to the normal direction of a fitting plane and the projection direction of a transverse axis on the plane, obtaining the local three-dimensional coordinate system, transforming the local displacement field data to the local three-dimensional coordinate system, and solving local strain according to a geometric relationship.

Has the advantages that:

compared with the prior art, the invention has the following advantages:

the traditional three-dimensional digital image correlation is expected to be greatly improved on the basis of the existing precision, and is difficult to realize at present no matter depending on the progress of an algorithm or hardware. Compared with the traditional three-dimensional digital image correlation method, the method has the advantages that the spatial resolution of the three-dimensional digital image correlation measurement is effectively improved by adopting a combination mode of the left and right three-dimensional camera arrays; the digital coding pattern is used for completing high-precision simultaneous calibration of internal and external parameters of the camera, so that the problems of complicated calibration and external parameter calibration between non-overlapping cameras are solved; meanwhile, a three-dimensional array parallel matching strategy based on polar lines and deformation constraints is used, so that the matching speed between the camera pictures is greatly increased, and efficient and high-precision non-contact three-dimensional full-field deformation measurement of materials and structures is realized.

Drawings

FIG. 1 is a flow chart of the inventive method.

Fig. 2 is a schematic diagram of a 3 × 3 three-dimensional camera array arrangement.

FIG. 3 is a schematic view of a digitally encoded pattern: (a) marker point encoding (b) digital speckle encoding.

Detailed Description

The invention is further described below with reference to the 3 x 3 array example and the accompanying drawings of the specification.

The invention relates to an array type high-resolution three-dimensional digital image correlation method for high-precision three-dimensional deformation measurement, which comprises the following steps of:

1) the processing and installation of two groups of 3X 3 two-dimensional camera arrays are completed, and a camera array arrangement method considering sample out-of-plane deformation is formulated at the same time, so that the data continuity of an initial state is guaranteed, the state continuity in the measuring process can be guaranteed, and the condition that the overlapped areas between the cameras are too small or even not overlapped can not occur.

For an array with a working distance Z, when the object is displaced from the plane by Δ Z, the following relationship can be obtained according to the imaging geometry:

where x, y are the image coordinates, Δ x is ImageSize-OverlapSize, ImageSize is the lateral dimension of the image, OverlapSize is the lateral dimension of the overlap region, and the subscripts 0 and 1 represent two adjacent cameras in the array, respectively. Should be used for ensuring data continuity

Is less thanThe initial OverlapSize is that an initial overlapping area is set according to the estimated maximum out-of-plane displacement, so that the condition that the camera overlapping area is too small or even does not overlap due to large out-of-plane displacement is avoided.

2) A three-dimensional camera array digital image related measurement system composed of two groups of two-dimensional camera arrays, a lighting device and a part to be measured is set up on an experimental site according to the requirements of a measurement view field and a working distance, and is shown in figure 2.

3) And preparing a random speckle pattern on the surface of the sample to be tested, and using the prepared speckle pattern as a position information carrier.

4) The calibration of the intrinsic parameters of each camera and the relative extrinsic parameters between cameras is accomplished using the code calibration board shown in fig. 3. Wherein, the internal reference calibration parameters comprise: principal point location (u)₀,v₀) Lens focal length f, pixel size (S)_x,S_y) And lens distortion parameter matrix D_iWherein the principal point position is the pixel coordinate of the intersection point of the optical axis of the lens and the target surface, and the lens distortion parameter matrix generally comprises 6-order radial distortion parameters K₁、K₂、K₃、K₄、K₅、K₆And 2 tangential distortion number P₁、P₂. Wherein, the external reference calibration parameters comprise: relative external parameters of the cameras and cameras in the array and camera external parameters of the cameras and cameras in the array, wherein each camera external parameter comprises 6 parameters n_x、n_y、n_z、t_x、t_y、t_z. The internal parameters of the cameras in the three-dimensional camera array can be directly calibrated by using a plane calibration method. The external parameters of the cameras in the three-dimensional camera array need to calibrate the relation between a camera coordinate system and a calibration board world coordinate system, and then the external parameters of the cameras in the array and the external parameters of the cameras among the array are solved according to the transformation of the coordinate system:

establishing a world coordinate system on a coding calibration board, wherein an external reference relation exists between the camera coordinate system of each camera and the world coordinate system, and the external reference relation comprises 6 parameters n_x、n_y、n_z、t_x、t_y、t_zThe rotation vector is transformed by Rodrigues to obtain a rotation matrix

According to the external reference relationship between one camera coordinate system C1 and the world coordinate system W and the external reference relationship between the other camera coordinate system C2 and the world coordinate system W, the external reference relationship between the two camera coordinate systems C1 and C2 in the three-dimensional camera array can be obtained:

5) adjusting the observation position and the observation angle of the three-dimensional camera array, ensuring that the detected area is completely imaged in the three-dimensional camera array, setting the cameras into a hard synchronization mode, and transmitting a synchronization signal through a signal trigger to enable the left and right groups of array cameras to simultaneously acquire sample images;

6) and (5) dividing the image collected in the step (5) into sub-regions, wherein the shape of the sub-regions adopts a polygon. Based on the two-dimensional digital image correlation technology, digital image correlation matching is carried out on sub-areas divided in the left image and the right image, the calculation efficiency is improved by adopting a parallel calculation method according to the polar line and deformation constraint matching principle, time sequence matching and three-dimensional matching between images of eighteen cameras are completed, and the right array image is matched and corresponding to the left array image. The matching algorithm in the digital image correlation matching selects a reverse combination Gauss-Newton algorithm, the shape function of the digital image correlation time sequence matching selects a first-order shape function, and the shape function of the digital image correlation stereo matching selects a second-order shape function;

7) and (4) according to the camera internal parameters and the inter-array camera external parameters obtained by calculation in the step (4), performing calculation on the point cloud data before and after deformation on the three-dimensional point cloud data corresponding to the left array camera image obtained by performing three-dimensional reconstruction matching in the step (6), and further obtaining three-dimensional displacement field data.

8) For point cloud data obtained by matching and reconstructing images of different cameras in the array, firstly converting the point cloud data into the same coordinate system according to the camera external parameters calibrated in the step (4); then, high-precision correction and splicing fusion are realized according to the spatial continuity of the point cloud data, and a splicing fusion mapping relation is obtained; on the basis of the splicing fusion mapping relation, high-precision splicing fusion of displacement field data is realized; and finally, realizing high-precision strain calculation of the local surface according to the three-dimensional topography data and the displacement field data.

When strain is calculated, firstly, plane fitting is carried out on local point cloud data, and a local three-dimensional coordinate system is determined according to the normal direction of a fitting plane and the projection direction of a transverse axis on the plane. And after a local three-dimensional coordinate system is obtained, transforming the local displacement field data to the local three-dimensional coordinate system and solving the full-field strain according to the geometric relation.

The above description is only a preferred embodiment of the present invention, and the scope of the present invention is not limited to the above embodiment, but equivalent modifications or changes made by those skilled in the art according to the present disclosure should be included in the scope of the present invention as set forth in the appended claims.