A calibration method for a 360-degree panoramic three-dimensional scanning measurement device

Abstract

This invention discloses a calibration method for a 360-degree panoramic 3D scanning measurement device. Based on the spatial structure of a conical surface, the calibration process is simple and highly accurate, enabling rapid and effective calibration of the 360-degree panoramic 3D scanning measurement device and achieving high-precision 3D panoramic measurement. The proposed calibration method features a simple calibration process, easy data acquisition, and is easy to calibrate and post-process. Placing a checkerboard pattern at different positions in the field of view of the panoramic ring camera allows for high-precision full-field calibration of the panoramic ring camera parameters. By calculating the 3D spatial plane equation of the checkerboard plane in the camera coordinate system for each checkerboard image and simultaneously solving the direction vector of the reflected light rays in the camera coordinate system, the input data for fitting the light cone surface can be accurately obtained. By introducing the light cone surface model, the objective optimization function of the light cone surface can be simplified, providing a better-performing optimization objective for subsequent nonlinear optimization algorithms, resulting in more accurate optimization results.

Description

Technical Field

[0001] This invention belongs to the field of three-dimensional measurement, specifically, it is a calibration method for a 360-degree panoramic three-dimensional scanning measurement device. Background Technology

[0002] 360° panoramic stereo measurement boasts a large field of view and is widely used in applications such as robot navigation, rail transit, assisted driving, pipeline measurement, and real-time scene positioning and measurement. Panoramic stereo scanning technology based on panoramic loop cameras features a compact structure, allowing a single camera to complete 360° panoramic 3D reconstruction. However, existing technologies suffer from the following technical problems: the measurement accuracy of the device depends on the calibration accuracy, and existing calibration methods are complex and difficult to use; furthermore, existing calibration methods based on device parameter structures have low accuracy. Summary of the Invention

[0003] The purpose of this invention is to address the shortcomings of existing technologies by proposing a calibration method for a 360-degree panoramic three-dimensional scanning measurement device. The calibration method proposed in this invention is based on the spatial structure of a conical surface, and the calibration process is simple and highly accurate.

[0004] This invention is achieved through the following technical solution:

[0005] A calibration method for a 360-degree panoramic three-dimensional scanning measurement device, wherein the calibration method is implemented by a calibration device, the calibration device comprising:

[0006] A panoramic ring camera consists of a camera and a panoramic ring lens mounted in front of the camera, used to obtain images with modulated circular structured light stripes.

[0007] A ring-shaped structured light generator is composed of a point laser generator and a coaxial axial cone mirror, which is used to generate ring-shaped structured light with a certain cone angle.

[0008] A quadric surface mirror is used to reflect the ring-shaped structured light generated by a ring-shaped structured light generator.

[0009] A computer is used to acquire images captured by the camera and to calibrate the device using calibration methods.

[0010] The panoramic ring camera, the quadric surface reflector, and the circular structured light generator are placed coaxially in order from top to bottom, and the computer is connected to the panoramic ring camera.

[0011] Calibration methods include:

[0012] A circular structured light is emitted by a circular structured light generator, and after being reflected by a checkerboard pattern, it is imaged onto a panoramic ring camera. The panoramic ring camera then captures multiple checkerboard images with circular structured light stripes.

[0013] The pixel coordinates of the corner points of the chessboard in the chessboard image are extracted using an image algorithm, and the pixel coordinates of the center line of the circular structured light stripes in the chessboard image are extracted using a center line algorithm.

[0014] Based on the panoramic ring camera calibration algorithm, the calibration parameters of the panoramic ring camera are obtained by optimizing the pixel coordinates of the checkerboard corner points;

[0015] Based on the calibration parameters of the panoramic ring camera, calculate the three-dimensional spatial plane equation of the chessboard plane in the camera coordinate system for each chessboard image.

[0016] Based on the calibration parameters of the panoramic ring camera, the center line pixel coordinates of the circular structured light stripe are transformed into the homogeneous coordinates of the circular structured light stripe in the camera coordinate system.

[0017] Based on the homogeneous coordinates of the circular structured light stripes, the equation of the corresponding three-dimensional straight line of the reflected ray in the camera coordinate system is calculated.

[0018] Based on the checkerboard three-dimensional spatial plane equation in the camera coordinate system and the three-dimensional spatial line equation of the reflected light in the camera coordinate system, the three-dimensional coordinates of the circular structured light stripe in the camera coordinate system are calculated.

[0019] Based on the model of the light cone, a target optimization function for the light cone is established, and the three-dimensional coordinates of the annular structured light stripe in the camera coordinate system are fitted to achieve the calibration of the light cone parameters of the device.

[0020] As a further improvement, the panoramic ring camera of the present invention acquires multiple checkerboard images with circular structured light stripes, and the checkerboard pattern needs to be evenly placed at different positions in the field of view of the panoramic ring camera.

[0021] As a further improvement, the panoramic ring camera calibration algorithm specifically employs the Mei model, using the checkerboard corner pixel coordinates to optimize and obtain the intrinsic parameters K, distortion D, mirror parameters ξ, and extrinsic parameters R of the panoramic ring camera. i and T i .

[0022] As a further improvement, the three-dimensional spatial plane equation of the chessboard image plane in the camera coordinate system of the present invention is as follows:

[0023] a i x+b i y+c i z+d i =0;

[0024] In the above formula, [a i ,b i ,c i ] is the normal vector of the plane of each chessboard square, which can be specifically represented by R.i Calculated

[0025] [a i ,b i ,c i ] T =R i [0,0,1] T ;

[0026] d i The position of each chessboard square can be determined by T. i Calculated

[0027]

[0028] As a further improvement, the center line pixel coordinates of the annular structured light stripes described in this invention are: Homogeneous coordinates of the annular structured light stripes in the camera coordinate system Transformation described as

[0029]

[0030] The corresponding equation of the reflected ray in the three-dimensional space of the camera coordinate system is described as follows:

[0031]

[0032] As a further improvement, the three-dimensional coordinates of the annular structured light stripe in the camera coordinate system are (x... j ,y j ,z j The objective function for the light cone surface is expressed as follows:

[0033]

[0034] In the above formula, [v x ,v y ,v z [x0, y0, z0] is the unit direction vector of the axis of the light cone, [x0, y0, z0] is the coordinate of the vertex of the light cone, and θ is the cone angle of the conic surface.

[0035] As a further improvement, the optical cone model described in this invention utilizes the geometric relationship of the optical cone generated by a circular ring structured light generator. The parameters of the optical cone can be described by a direction vector, a vertex, and a cone angle. Specifically, the equation of the optical cone in the panoramic ring camera coordinate system is described as follows:

[0036]

[0037] In the above formula, [v x ,v y ,vz [x0, y0, z0] is the unit direction vector of the axis of the light cone, [x0, y0, z0] is the coordinate of the vertex of the light cone, and θ is the cone angle of the conic surface.

[0038] The beneficial effects of this invention are as follows:

[0039] 1. A calibration method for a ring structure based on an optical cone surface model is proposed, which can quickly and effectively calibrate a 360-degree panoramic three-dimensional scanning measurement device and achieve high-precision three-dimensional panoramic measurement.

[0040] 2. The proposed calibration method has a simple calibration process, simple data acquisition, and is easy to calibrate and post-process.

[0041] 3. Placing the checkerboard pattern at different positions within the panoramic loop camera's field of view enables high-precision full-field calibration of the panoramic loop camera's parameters.

[0042] 4. By calculating the three-dimensional spatial plane equation of the checkerboard plane in the camera coordinate system for each checkerboard image, and simultaneously solving the direction vector of the reflected light rays in the camera coordinate system, the input data for fitting the light cone surface can be accurately obtained.

[0043] 5. By introducing the model of the light conic surface, the objective optimization function of the light conic surface can be simplified, providing a better optimization objective for subsequent nonlinear optimization algorithms, thus making the optimization results more accurate. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of a 360-degree panoramic 3D scanning measurement device;

[0045] Figure 2 This is an optical structure diagram of a ring-shaped light generator;

[0046] Figure 3 This is a flowchart of the device calibration method;

[0047] Figure 1 In China: 1. Camera; 2. Panoramic ring lens; 3. Quadratic surface reflector; 4. Circular structured light generator; 5. Computer;

[0048] Figure 2 In the middle: 3. Quadratic surface mirror 6. Axicon mirror 7. Point laser generator. Detailed Implementation

[0049] The technical solution of the present invention will be further described below with reference to the accompanying drawings and specific embodiments:

[0050] Example 1:

[0051] This invention proposes a calibration method for a 360-degree panoramic three-dimensional scanning measurement device. The calibration method is implemented using a calibration device. Figure 1 This is a schematic diagram of a 360-degree panoramic 3D scanning measurement device; the calibration device includes:

[0052] A panoramic ring-shaped camera, consisting of camera 1 and a panoramic ring-shaped lens 2 mounted in front of the camera, is used to obtain images with modulated circular structured light stripes; a circular structured light generator 4, consisting of a point laser generator 7 and a coaxial cone mirror 6, is used to obtain images with modulated circular structured light stripes. Figure 2 This is an optical structure diagram of a circular structured light generator; used to generate circular structured light with a certain cone angle; a quadric surface mirror 3, used to reflect the circular structured light generated by the circular structured light generator; a computer 5, used to acquire images captured by the camera and to complete the calibration of the device through a calibration method; the panoramic ring camera, quadric surface mirror 3, and circular structured light generator 4 are placed coaxially in order from top to bottom, and the computer 5 is connected to the panoramic ring camera;

[0053] Calibration principle of the present invention

[0054] A panoramic ring camera is a modified single-viewpoint catadioptric optical device. The general projection model of the device is as follows: at a point P on the light cone surface... w Point P in the camera 1 coordinate system is transformed from the world coordinate system by the extrinsic parameter matrices (R and T). m Transformed into the normalized plane π through mirror transformation n Point P on n The specific conversion process can be described as follows:

[0055] (a) Point P in the coordinate system of camera 1 m Point P projected onto the unit sphere s =(X s ,Y s Z s ),

[0056] (b) These points are then transformed into a new coordinate system, the origin of which is C. m = (0,0,0) becomes C p = (0,0,ξ), therefore P s =(X s ,Y s Z s +ξ)

[0057] (c) The point is then projected onto the normalized plane.

[0058] This mirror transformation process is denoted as H.

[0059] Finally P n Transformed to pixel plane π using intrinsic parameter matrix (K) and distortion coefficients (D). p Point P onp Therefore, the entire transformation process can be described as follows:

[0060] G=[KD]H[RT]#(1)

[0061] A circular structured light is a diverging conical surface, which can be described in the camera-1 coordinate system as...

[0062]

[0063] In the above formula, [v x ,v y ,v z [x0, y0, z0] is the unit direction vector of the axis of the light cone, [x0, y0, z0] is the coordinate of the vertex of the light cone, and θ is the cone angle of the conic surface.

[0064] The calibration accuracy of the optical cone directly affects the final measurement accuracy of the device. The panoramic ring camera is a typical single-viewpoint catadioptric optical device, and the calibration methods for such single-viewpoint catadioptric optical devices have been extensively studied.

[0065] This invention discloses a calibration method for a 360-degree panoramic three-dimensional scanning measurement device, comprising the following steps: Figure 3 The flowchart for the device calibration method is described in detail below:

[0066] Step 1: A circular structured light is emitted through the circular structured light generator 4. After being reflected sequentially by the quadratic surface emitting mirror 3 and the checkerboard pattern, the light is imaged onto the panoramic ring-shaped camera. The panoramic ring-shaped camera captures multiple checkerboard images with circular structured light stripes. The checkerboard pattern needs to be evenly placed at different positions within the panoramic ring-shaped camera's field of view.

[0067] Step 2: Extract the pixel coordinates of the corner points of the checkerboard in the checkerboard image using an image algorithm, and extract the pixel coordinates of the center line of the circular structured light stripes in the checkerboard image using a center line algorithm;

[0068] Step 3: Based on the panoramic ring camera calibration algorithm, the calibration parameters of the panoramic ring camera are obtained by optimizing the pixel coordinates of the checkerboard corner points. Specifically, the panoramic ring camera calibration algorithm uses the Mei model to optimize the intrinsic parameters K, distortion D, mirror parameters ξ, and extrinsic parameters R of the panoramic ring camera using the pixel coordinates of the checkerboard corner points. i and T i .

[0069] Step 4: Based on the calibration parameters of the panoramic loop camera, calculate the three-dimensional spatial plane equation of the checkerboard plane in the camera coordinate system for each checkerboard image; the three-dimensional spatial plane equation of the checkerboard plane in the camera coordinate system for each checkerboard image is as follows:

[0070] a i x+b i y+c i z+d i =0;

[0071] In the above formula, [a i ,b i ,c i ] is the normal vector of the plane of each chessboard square, which can be specifically represented by R. i Calculated

[0072] [a i ,b i ,c i ] T =R i [0,0,1] T ;

[0073] d i The position of each chessboard square can be determined by T. i Calculated

[0074]

[0075] Step 5: Based on the calibration parameters of the panoramic ring camera, transform the center line pixel coordinates of the circular structured light stripes into homogeneous coordinates of the circular structured light stripes in the camera coordinate system.

[0076] Step 6: Based on the homogeneous coordinates of the circular structured light stripes, calculate the corresponding three-dimensional spatial equation of the reflected ray in the camera coordinate system.

[0077] The center line pixel coordinates of the circular structured light stripe are Homogeneous coordinates of the annular structured light stripes in the camera coordinate system Transformation described as

[0078]

[0079] The corresponding equation of the reflected ray in the three-dimensional space of the camera coordinate system is described as follows:

[0080]

[0081] Step 7: Based on the checkerboard 3D spatial plane equation in the camera coordinate system and the reflected ray 3D spatial line equation in the camera coordinate system, calculate the 3D coordinates of the circular structured light stripe in the camera coordinate system; the 3D coordinates of the circular structured light stripe in the camera coordinate system are (x... j ,y j ,z j );

[0082] Step 8: Based on the model of the light cone, establish the objective optimization function for the light cone, fit the three-dimensional coordinates of the annular structured light stripes in the camera coordinate system, and calibrate the parameters of the light cone of the device. The objective optimization function for the light cone is expressed as follows:

[0083]

[0084] In the above formula, [v x ,v y ,v z [x0, y0, z0] is the unit direction vector of the optical cone axis, [x0, y0, z0] is the coordinate of the vertex of the optical cone, and θ is the cone angle of the conic surface. The fitting method is to obtain the optical cone parameters through the Levenberg-Marquardt nonlinear optimization algorithm.

[0085] The light cone model utilizes the geometric relationships of the light cone generated by a circular structured light generator. The parameters of the light cone can be described by a direction vector, a vertex, and a cone angle. Specifically, the equation of the light cone in the panoramic ring camera coordinate system is described as follows:

[0086]

[0087] In the above formula, [v x ,v y ,v z [x0, y0, z0] is the unit direction vector of the axis of the light cone, [x0, y0, z0] is the coordinate of the vertex of the light cone, and θ is the cone angle of the conic surface.

[0088] After the device is calibrated, it performs three-dimensional measurements based on the calibrated light cone parameters and the modulation image of the circular structured light on the object surface.

[0089] The above is not intended to limit the present invention. It should be noted that, for those skilled in the art, various changes, modifications, additions or substitutions can be made without departing from the essential scope of the present invention, and these improvements and refinements should also be considered within the scope of protection of the present invention.

Claims

1. A calibration method for a 360-degree panoramic three-dimensional scanning measurement device, characterized in that the calibration method is implemented through a calibration device, the calibration device comprising: A panoramic ring camera consists of a camera and a panoramic ring lens mounted in front of the camera, used to obtain images with modulated circular structured light stripes. A ring-shaped structured light generator is composed of a point laser generator and a coaxial axial cone mirror, which is used to generate ring-shaped structured light with a certain cone angle. A quadric surface mirror is used to reflect the ring-shaped structured light generated by a ring-shaped structured light generator. A computer is used to acquire images captured by the camera and to calibrate the device using calibration methods. The panoramic ring camera, the quadrature surface reflector, and the circular structured light generator are placed coaxially in order from top to bottom, and the computer is connected to the panoramic ring camera. The calibration method includes: A circular structured light is emitted by a circular structured light generator, and after being reflected by a checkerboard pattern, it is imaged onto a panoramic ring camera. The panoramic ring camera then captures multiple checkerboard images with circular structured light stripes. The pixel coordinates of the corner points of the chessboard in the chessboard image are extracted using an image algorithm, and the pixel coordinates of the center line of the circular structured light stripes in the chessboard image are extracted using a center line algorithm. Based on the panoramic ring camera calibration algorithm, the calibration parameters of the panoramic ring camera are obtained by optimizing the pixel coordinates of the checkerboard corner points; Based on the calibration parameters of the panoramic ring camera, calculate the three-dimensional spatial plane equation of the chessboard plane in the camera coordinate system for each chessboard image. Based on the calibration parameters of the panoramic ring camera, the center line pixel coordinates of the circular structured light stripe are transformed into the homogeneous coordinates of the circular structured light stripe in the camera coordinate system. Based on the homogeneous coordinates of the circular structured light stripes, the equation of the corresponding three-dimensional straight line of the reflected ray in the camera coordinate system is calculated. Based on the checkerboard three-dimensional spatial plane equation in the camera coordinate system and the three-dimensional spatial line equation of the reflected light in the camera coordinate system, the three-dimensional coordinates of the circular structured light stripe in the camera coordinate system are calculated. Based on the model of the light cone, a target optimization function for the light cone is established, and the three-dimensional coordinates of the annular structured light stripe in the camera coordinate system are fitted to achieve the calibration of the light cone parameters of the device.

2. The calibration method for the 360-degree panoramic three-dimensional scanning measurement device according to claim 1, characterized in that, The panoramic ring camera captures multiple checkerboard images with circular structured light stripes. The checkerboard pattern needs to be evenly placed at different positions within the panoramic ring camera's field of view.

3. The calibration method for the 360-degree panoramic three-dimensional scanning measurement device according to claim 1 or 2, characterized in that, The aforementioned panoramic ring camera calibration algorithm specifically employs the Mei model, optimizing the intrinsic parameters K, distortion D, mirror parameters ξ, and extrinsic parameters R of the panoramic ring camera using the pixel coordinates of the checkerboard corner points. i and T i .

4. The calibration method for the 360-degree panoramic three-dimensional scanning measurement device according to claim 3, characterized in that, The three-dimensional spatial plane equation of each chessboard image in the camera coordinate system is as follows: a i x+b i y+c i z+d i =0; In the above formula, [a i ,b i ,c i ] is the normal vector of the plane of each chessboard square, which can be specifically represented by R. i Calculated [a i ,b i ,c i ] T =R i [0,0,1] T ; d i The position of each chessboard square can be determined by T. i Calculated 5. The calibration method for the 360-degree panoramic three-dimensional scanning measurement device according to claim 3, characterized in that, The center line pixel coordinates of the circular structured light stripe are Homogeneous coordinates of the annular structured light stripes in the camera coordinate system The transformation is described as follows The corresponding equation of the reflected ray in the three-dimensional space of the camera coordinate system is described as follows:

6. The calibration method for the 360-degree panoramic three-dimensional scanning measurement device according to claim 1, 2, 4, or 5, characterized in that, The three-dimensional coordinates of the circular structured light stripe in the camera coordinate system are (x... j ,y j ,z j The objective optimization function for the light cone surface is expressed as follows: In the above formula, [v x ,v y ,v z [x0, y0, z0] is the unit direction vector of the axis of the light cone, [x0, y0, z0] is the coordinate of the vertex of the light cone, and θ is the cone angle of the conic surface; The fitting method described above involves obtaining the parameters of the optical cone surface using the Levenberg-Marquardt nonlinear optimization algorithm.

7. The calibration method for the 360-degree panoramic three-dimensional scanning measurement device according to claim 6, characterized in that, The aforementioned light cone model utilizes the geometric relationships of the light cone generated by a circular structured light generator. The parameters of the light cone can be described by a direction vector, a vertex, and a cone angle. Specifically, the equation of the light cone in the panoramic ring camera coordinate system is described as follows: In the above formula, [v x ,v y ,v z [x0, y0, z0] is the unit direction vector of the axis of the light cone, [x0, y0, z0] is the coordinate of the vertex of the light cone, and θ is the cone angle of the conic surface.

Images