Omnibearing high-precision three-dimensional measurement system and method

Abstract

The invention discloses an omnibearing high-precision three-dimensional measurement system and method, which can realize omnibearing measurement of a target object and is suitable for high-precision three-dimensional reconstruction of various objects with complex shapes. A to-be-measured target is placed on the rotary table, the three-dimensional camera is placed on the pitching cradle head, the pitching cradle head is installed on the lifting guide rail, and the motion platform carries out data acquisition on the to-be-measured target under multiple shooting view angles; the upper computer controls the rotary table to drive a to-be-measured target to rotate, and the lifting guide rail and the pitching holder drive the three-dimensional camera to lift and pitch; in the measurement process, the upper computer controls the three-dimensional camera to select an optimal measurement visual angle to measure a to-be-measured target according to a measurement strategy, a result with a poor measurement effect is measured repeatedly, and the measurement result is complemented; in the calibration process, the calibration plate is placed on the rotary table, the transverse error, the reflectivity error and the distance error of the three-dimensional camera are calibrated respectively, and measurement errors are eliminated.

Description

Technical Field

[0001] The present invention belongs to the technical field of three-dimensional imaging systems, and in particular relates to an omnidirectional high-precision three-dimensional measurement system and an omnidirectional high-precision three-dimensional measurement method. Background Art

[0002] Three-dimensional measurement technology has been widely used in industrial manufacturing, cultural heritage protection, architectural measurement and other fields. However, existing three-dimensional measurement systems usually rely on a fixed single perspective for data collection, which results in limited viewing angles during the measurement process and makes it difficult to achieve all-round measurement of the target object. At the same time, in the multi-perspective stitching process, the overall measurement accuracy is often affected due to the large data alignment error at different perspectives, especially when the target object has a complex structure or variable surface characteristics. The problem of stitching error is particularly prominent. Although some studies have proposed a method of using multiple cameras to collect data from different directions at the same time, this method has problems such as high hardware cost and high system complexity, and it is difficult to promote and apply in some scenarios that require flexible movement and high-precision measurement.

[0003] In addition, the existing 3D measurement systems mostly use local registration methods when stitching data, which is easily affected by the local data accuracy, resulting in inconsistency in overall data reconstruction. In omnidirectional 3D measurement, as the measurement range increases, the accumulation of local errors will be further amplified, resulting in larger stitching errors, which cannot meet the needs of high-precision applications.

[0004] Therefore, the existing three-dimensional measurement system still has many technical bottlenecks in the process of achieving high-precision all-round measurement. How to effectively improve the accuracy of the three-dimensional measurement system, reduce the multi-view stitching error, and simplify the system architecture has become a technical problem that needs to be solved urgently. Summary of the invention

[0005] In order to overcome the defects of the prior art, the technical problem to be solved by the present invention is to provide an omnidirectional high-precision 3D measurement system, which can overcome the limitations of existing fixed-viewing angle measurement, realize omnidirectional measurement of target objects, and is suitable for high-precision 3D reconstruction of objects of various complex shapes. It can significantly reduce the error accumulation of multi-viewing angle data and improve the overall accuracy of 3D measurement results. It has a flexible structure and is easy to operate. It can adjust the camera position and angle according to measurement requirements to ensure the stability and consistency of the measurement process. It is suitable for a variety of measurement environments and application scenarios.

[0006] The technical solution of the present invention is: this all-round high-precision three-dimensional measurement system comprises: a three-degree-of-freedom motion platform, a three-dimensional camera (7), a calibration plate, and a host computer (8);

[0007] The three-degree-of-freedom motion platform comprises: an X-direction truss (1), a Y-direction truss (2), a Z-direction truss (3), a turntable (4), an elevating guide rail (5), and a pitch platform (6). The target to be measured is placed on the turntable, the three-dimensional camera is placed on the pitch platform, and the pitch platform is installed on the elevating guide rail. The motion platform collects data of the target to be measured under multiple shooting angles.

[0008] The host computer controls the turntable to drive the target to be measured to rotate, and the lifting rail and the pitch pan / tilt table drive the 3D camera to lift and pitch;

[0009] During the 3D measurement process, the host computer controls the 3D camera to select the optimal measurement angle to measure the target according to the measurement strategy, and repeats the measurement results with poor results to complete the measurement results.

[0010] During the calibration process of the 3D measurement system, the calibration plate is placed on a turntable to calibrate the lateral error, reflectivity error, and distance error of the 3D camera to eliminate the measurement error of the 3D measurement system.

[0011] Compared with the prior art, the present invention has the following beneficial effects:

[0012] 1. All-round measurement: The design of the three-degree-of-freedom motion platform can achieve all-round measurement of the target object. The combination of the turntable, lifting rail and tilting pan / tilt platform enables the 3D camera to collect data from multiple perspectives, overcoming the limitations of existing fixed-perspective measurement and is suitable for high-precision 3D reconstruction of objects of various complex shapes.

[0013] 2. High-precision stitching: The global consistency algorithm is used to optimize the frame-by-frame registration. By calibrating the 3D imaging process, the error in stitching data from different perspectives is eliminated. Compared with the traditional local registration method, it can significantly reduce the error accumulation of multi-perspective data and improve the overall accuracy of the 3D measurement results.

[0014] 3. Flexibility and stability: The three-degree-of-freedom motion platform used in the system has a flexible structure and is easy to operate. It can adjust the camera position and angle according to measurement requirements. At the same time, through the combination of calibration and optimization algorithms, the stability and consistency of the measurement process are guaranteed, which is suitable for a variety of measurement environments and application scenarios.

[0015] A working method of the omnidirectional high-precision three-dimensional measurement system is also provided, which comprises the following steps:

[0016] (1) Calibrate the 3D measurement system to eliminate lateral error and reflectivity error;

[0017] (2) In the point cloud frame-by-frame registration stage, the consistency optimization method is used to eliminate the accumulated errors in the stitching process. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a schematic diagram of the circular calibration plate in the present invention.

[0019] Figure 2 It is a schematic diagram of the reflectivity calibration plate in the present invention.

[0020] Figure 3 It is a schematic diagram of the calibration process in the present invention.

[0021] Figure 4 It is a flow chart of the calibration process in the present invention.

[0022] Figure 5 It is a schematic structural diagram of the omnidirectional high-precision three-dimensional measurement system of the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0024] In order to make the description of the present disclosure more detailed and complete, the following provides an illustrative description of the implementation and specific embodiments of the present invention; however, this is not the only form of implementing or using the specific embodiments of the present invention. The implementation covers the features of multiple specific embodiments and the method steps and sequences used to construct and operate these specific embodiments. However, other specific embodiments can also be used to achieve the same or equivalent functions and step sequences.

[0025] like Figure 5 As shown, this all-round high-precision three-dimensional measurement system includes: a three-degree-of-freedom motion platform, a three-dimensional camera 7, a calibration plate, and a host computer 8;

[0026] The three-degree-of-freedom motion platform includes: an X-direction truss 1, a Y-direction truss 2, a Z-direction truss 3, a turntable 4, a lifting guide rail 5, and a pitch platform 6. The target to be measured is placed on the turntable, the three-dimensional camera is placed on the pitch platform, and the pitch platform is installed on the lifting guide rail. The motion platform collects data of the target to be measured under multiple shooting angles;

[0027] The host computer controls the turntable to drive the target to be measured to rotate, and the lifting rail and the pitch pan / tilt table drive the 3D camera to lift and pitch;

[0028] During the 3D measurement process, the host computer controls the 3D camera to select the optimal measurement angle to measure the target according to the measurement strategy, and repeats the measurement results with poor results to complete the measurement results.

[0029] During the calibration process of the 3D measurement system, the calibration plate is placed on a turntable to calibrate the lateral error, reflectivity error, and distance error of the 3D camera to eliminate the measurement error of the 3D measurement system.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] 1. All-round measurement: The design of the three-degree-of-freedom motion platform can achieve all-round measurement of the target object. The combination of the turntable, lifting rail and tilting pan / tilt platform enables the 3D camera to collect data from multiple perspectives, overcoming the limitations of existing fixed-perspective measurement and is suitable for high-precision 3D reconstruction of objects of various complex shapes.

[0032] 2. High-precision stitching: The global consistency algorithm is used to optimize the frame-by-frame registration. By calibrating the 3D imaging process, the error in stitching data from different perspectives is eliminated. Compared with the traditional local registration method, it can significantly reduce the error accumulation of multi-perspective data and improve the overall accuracy of the 3D measurement results.

[0033] 3. Flexibility and stability: The three-degree-of-freedom motion platform used in the system has a flexible structure and is easy to operate. It can adjust the camera position and angle according to measurement requirements. At the same time, through the combination of calibration and optimization algorithms, the stability and consistency of the measurement process are guaranteed, which is suitable for a variety of measurement environments and application scenarios.

[0034] Preferably, the calibration plate comprises: a circular plate (such as Figure 1 As shown), reflectivity plate (as Figure 2 As shown); there are M×N circular protrusions on the surface of the circular calibration plate; there are i×i circular protrusions evenly distributed on the surface of the reflectivity calibration plate, and the reflectivity decreases from left to right.

[0035] A working method of the omnidirectional high-precision three-dimensional measurement system is also provided, which comprises the following steps:

[0036] (1) Calibrate the 3D measurement system to eliminate lateral error and reflectivity error;

[0037] (2) In the point cloud frame-by-frame registration stage, the consistency optimization method is used to eliminate the accumulated errors in the stitching process.

[0038] Preferably, if Figure 3 , 4 As shown, the step (1) includes the following sub-steps:

[0039] (1.1) Place the circular calibration plate on the turntable and adjust the lifting rails and the tilting platform to make the circular calibration plate parallel to the imaging plane of the 3D camera;

[0040] (1.2) Adjust the distance between the circular calibration plate and the imaging system to {D1,...,D L}, respectively collect the original point cloud {P1,P2,...,P L}, and calculate the standard circle center matrix set T{T1,T2,...,T L};

[0041] (1.3) For each distance of the original point cloud P i , use Euclidean distance segmentation to obtain the sub-point cloud set Q;

[0042] (1.4) For the sub-point cloud Q in the sub-point cloud set Q j ∈Q, use RANSAC to identify and get the center O j (x j ,y j ), add all the calculated circle centers to the circle center set R i middle;

[0043] (1.5) Sparse calibration matrix calculation LUT i * =R i . / T i , the sparse calibration matrix is ​​transformed along the x direction (such as Figure 5 The x-direction of the truss 1) and the y-direction (such as Figure 5 The Y direction of the truss 2) is decomposed into

[0044] (1.6) Use the thin plate spline interpolation algorithm to densify the sparse calibration matrix and obtain the LUT x ,LUT y , thus obtaining the calibration matrix set at all distances

[0045] LUT={LUT1,...LUT L};

[0046] (1.7) Use the calibration matrix to calibrate the collected point cloud C Raw To calibrate to eliminate distortion error, first calculate C Raw Mean C , judge the Mean C Belong to the interval

[0047] [D i ,D i+1 ], execute formula (1) to complete the calibration:

[0048] C=C Raw .*LUT i-x .*LUT i-y (1);

[0049] (1.8) Replace the circular calibration plate with a reflectivity plate and repeat steps (1) to (7).

[0050] Preferably, the step (2) comprises the following sub-steps:

[0051] (2.1) Record the initial shooting posture of the 3D measurement system, move the lifting rail and the tilting platform quantitatively in sequence, and record the posture set A of each shooting angle during the movement;

[0052] (2.2) Match the poses of two adjacent frames of point clouds respectively, transform all point cloud sets into the world coordinate system, obtain the transformed pose C of two adjacent frames of point clouds, and obtain the local constraint relationship pose set B;

[0053] (2.3) The common view area in the pose set A is the global boundary, and the common view area in the two adjacent frame poses C is the local boundary. A consistent objective function is constructed to minimize the error between the global boundary and the local boundary to complete the point cloud registration.

[0054] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any simple modification, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention are still within the protection scope of the technical solution of the present invention.

Claims

1. An all-round high-precision three-dimensional measurement system, characterized in that: It includes: Three-degree-of-freedom motion platform, three-dimensional camera (7), calibration board, host computer (8); The three-degree-of-freedom motion platform includes: an X-direction truss (1), a Y-direction truss (2), a Z A direction truss (3), a turntable (4), a lifting guide rail (5), and a pitch platform (6), wherein the target to be measured is placed on the turntable, the three-dimensional camera is placed on the pitch platform, the pitch platform is mounted on the lifting guide rail, and the motion platform collects data on the target to be measured under multiple shooting angles; The host computer controls the turntable to drive the target to be measured to rotate, and the lifting rail and the pitch pan / tilt table drive the 3D camera to lift and pitch; During the 3D measurement process, the host computer controls the 3D camera to select the optimal measurement angle to measure the target according to the measurement strategy, and repeats the measurement results with poor results to complete the measurement results. During the calibration process of the 3D measurement system, the calibration plate is placed on a turntable to calibrate the lateral error, reflectivity error, and distance error of the 3D camera to eliminate the measurement error of the 3D measurement system.

2. The omnidirectional high-precision three-dimensional measurement system according to claim 1, characterized in that: The calibration plate comprises: a circular plate and a reflectivity plate; the surface of the circular calibration plate has M×N circular protrusions; the surface of the reflectivity calibration plate has i×i circular protrusions evenly distributed, and the reflectivity decreases from left to right.

3. The working method of the omnidirectional high-precision three-dimensional measurement system according to claim 1, characterized in that: The working method comprises the following steps: (1) Calibrate the 3D measurement system to eliminate lateral error and reflectivity error; (2) In the point cloud frame-by-frame registration stage, the consistency optimization method is used to eliminate the accumulated errors in the stitching process.

4. The working method of the omnidirectional high-precision three-dimensional measurement system according to claim 3, characterized in that: The step (1) comprises the following sub-steps: (1.1) Place the circular calibration plate on the turntable and adjust the lifting rails and the tilting platform to make the circular calibration plate parallel to the imaging plane of the 3D camera; (1.2) Adjust the distance between the circular calibration plate and the imaging system to {D1,...,D L }, respectively collect the original point cloud {P1,P2,...,P L }, and calculate the standard circle center matrix set T{T1,T2,...,T L }; (1.3) For each distance of the original point cloud P i , use Euclidean distance segmentation to obtain the sub-point cloud set Q; (1.4) For the sub-point cloud Q in the sub-point cloud set Q j ∈Q, use RANSAC to identify and get the center O j (x j ,y j ), add all the calculated circle centers to the circle center set R i middle; (1.5) Sparse calibration matrix calculation LUT i * =R i . / T i , decompose the sparse calibration matrix along the x and y directions to obtain (1.6) Use the thin plate spline interpolation algorithm to densify the sparse calibration matrix and obtain the LUT x ,LUT y , thus obtaining the calibration matrix set at all distances LOT={LOTS1,...LOTS L }; (1.7) Use the calibration matrix to calibrate the collected point cloud C Raw To calibrate to eliminate distortion error, first calculate C Raw Mean C , judge the Mean C The interval [D i ,D i+1 ], execute formula (1) to complete the calibration: C=C Raw .*LUT i-x .*LUT i-y (1); (1.8) Replace the circular calibration plate with a reflectivity plate and repeat steps (1) to (7).

5. The working method of the omnidirectional high-precision three-dimensional measurement system according to claim 4, characterized in that: The step (2) comprises the following sub-steps: (2.1) Record the initial shooting posture of the 3D measurement system, move the lifting rail and the tilting platform quantitatively in sequence, and record the posture set A of each shooting angle during the movement; (2.2) Match the poses of two adjacent frames of point clouds respectively, transform all point cloud sets into the world coordinate system, obtain the transformed pose C of two adjacent frames of point clouds, and obtain the local constraint relationship pose set B; (2.3) The common view area in the pose set A is the global boundary, and the common view area in the two adjacent frame poses C is the local boundary. A consistent objective function is constructed to minimize the error between the global boundary and the local boundary to complete the point cloud registration.

Images