Multi-camera based three-dimensional imaging method, system and computer readable storage medium

Abstract

The application discloses a kind of three-dimensional imaging method, system and computer readable storage medium based on multiple cameras.The method is applied to the three-dimensional imaging system including multiple cameras, comprising: using the three-dimensional imaging system of calibration takes target object;For the pixel point on the image that first camera takes, find the sub-pixel point on the image that other cameras take;For the pixel point on the image of first camera and the sub-pixel point on the image of other cameras, by normalization and distortion removal, obtain the projection point on the normalized imaging plane;For each group of projection points, from the principal point of each camera, multiple straight lines pointing to the projection point are drawn;For each group of straight lines, obtain the spatial point coordinates of the minimum Euclidean distance from each straight line in the group of straight lines in space, as the three-dimensional space coordinates of target object.The above-mentioned three-dimensional imaging method, system and computer readable storage medium can be more accurately positioned, improve the accuracy of three-dimensional reconstruction.

Description

Technical Field

[0001] This invention relates to the field of three-dimensional imaging technology, and in particular to a three-dimensional imaging method, system, and computer-readable storage medium based on multiple cameras. Background Technology

[0002] Fringe projection profilometry (FPP) is a commonly used optical 3D imaging technique using structured light active illumination. Its principle involves projecting sinusoidal fringes onto the surface of the object being measured, while simultaneously controlling a camera to acquire deformed fringes modulated by the object's height. Phase information is then obtained using fringe analysis methods, and finally, based on pre-calibrated system parameters, the 3D contour information of the object is reconstructed.

[0003] Traditional 3D reconstruction methods suffer from low accuracy. Summary of the Invention

[0004] Therefore, it is necessary to address the issue of low accuracy in traditional 3D reconstruction schemes by providing a multi-camera-based 3D imaging method, system, and computer-readable storage medium.

[0005] A multi-camera-based 3D imaging method, applied to a 3D imaging system including multiple cameras, the method comprising:

[0006] The target object is photographed using a calibrated 3D imaging system;

[0007] Using any one of the multiple cameras as the first camera, find the sub-pixel points in the images of the target object captured by the first camera that correspond to the pixels in the images of the target object captured by other cameras.

[0008] For the pixels in the image of the first camera and the sub-pixels in the images of other cameras, normalization and distortion correction are performed to obtain the projection points on the normalized imaging plane; among them, the projection points corresponding to the same point on the target object are a group.

[0009] For each set of projection points on the normalized imaging plane, multiple straight lines are drawn from the principal point of each camera, pointing to the corresponding camera projection point; among them, the straight lines drawn from the same set of projection points constitute a set.

[0010] For each set of straight lines, obtain the coordinates of the spatial point that minimizes the sum of the Euclidean distances from each straight line in that set, and use these coordinates as the reconstructed 3D spatial coordinates of the corresponding point on the target object.

[0011] In one embodiment, the method further includes calibrating the three-dimensional imaging system to obtain the intrinsic and extrinsic parameters of each camera.

[0012] In one embodiment, the normalization is performed on distorted pixels on the camera according to the following formula. The distortion-normalized coordinates are obtained by projecting the image onto the normalized imaging plane. :

[0013] ;

[0014] Where f is the focal length of the camera. The coordinates are the principal point of the camera, and the distorted pixels include pixels on the image captured by the first camera and sub-pixels on the images captured by other cameras.

[0015] In one embodiment, the distortion correction normalizes the distortion coordinates according to the following formula. Convert to distortion-free normalized coordinates :

[0016] ;

[0017] in, , The radial distortion coefficient is... denoted as the tangential distortion coefficient.

[0018] In one embodiment, the three-dimensional imaging system further includes a stripe projector, and the step of using the calibrated three-dimensional imaging system to photograph the target object includes: using the calibrated three-dimensional imaging system to photograph the target object on which orthogonal stripes are projected by the stripe projector;

[0019] The step of finding sub-pixel points on images of the target object captured by other cameras that correspond to pixels on the image of the first camera includes: the multiple cameras acquiring orthogonal fringes to obtain an orthogonal phase distribution map for each camera, and using an orthogonal phase analysis method to analyze the orthogonal phase distribution map to obtain sub-pixel points on images captured by other cameras that correspond to pixels on the image of the first camera.

[0020] In one embodiment, obtaining the coordinates of the spatial point in space that minimizes the sum of the Euclidean distances to each of the set of lines includes:

[0021] Obtain the expression for each line in the coordinate system of any j-th camera; where, N is the number of cameras;

[0022] Establish spatial points The expression for the distance to the j-th line The j-th straight line corresponds to the j-th camera;

[0023] Establish an expression for the sum of the Euclidean distances of each line in this set of lines:

[0024] ;

[0025] Regarding the The minimum value is calculated to obtain the coordinates of the required spatial point.

[0026] In one embodiment, the expression for each straight line in the coordinate system of the j-th camera is represented using point normal form as follows:

[0027] ;

[0028] in Represents the coordinates of a point on the j-th straight line. Let represent the normal vector of the j-th line corresponding to the i-th pixel. represents the coordinates of the principal point of the j-th camera; t represents the proportionality coefficient of the distance between a point on a straight line in space and the principal point of the camera.

[0029] The expression for the distance from a point in space to each straight line is:

[0030] ;

[0031] in:

[0032] .

[0033] In one embodiment, the Newton-Raphson iterative method is used to solve the problem.

[0034] A multi-camera-based 3D imaging system includes multiple cameras, a stripe projector, a memory, and a processor; the multiple cameras capture images of a target object onto which orthogonal stripes are projected by the stripe projector; the memory stores a computer program; and the processor executes the computer program to process the image of the target object, thereby implementing the steps of the method described above.

[0035] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0036] The aforementioned 3D imaging method, apparatus, and computer storage medium employ multiple cameras, and the coordinates of point M on the target object lie within an uncertain region where three or more conical regions intersect. Therefore, the range of the uncertain region is smaller, allowing for more precise positioning and improving the accuracy of 3D reconstruction. Attached Figure Description

[0037] Figure 1 This is a schematic diagram illustrating the composition and working principle of a binocular 3D imaging system.

[0038] Figure 2 A schematic diagram of the uncertain space where light rays from the left and right cameras intersect;

[0039] Figure 3 This is a flowchart of a multi-camera 3D imaging method according to one embodiment;

[0040] Figure 4 This is a flowchart illustrating the process of obtaining the coordinates of a spatial point in a multi-camera 3D imaging method, where the sum of the Euclidean distances to each of a set of straight lines is minimized.

[0041] Figure 5 This is a schematic diagram of an uncertain region in a multi-camera 3D imaging method according to an embodiment;

[0042] Figure 6A and Figure 6B This is a comparison of the reconstruction accuracy of a four-camera 3D imaging system and a two-camera system, as shown in this embodiment. Detailed Implementation

[0043] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Preferred embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0044] To illustrate the three-dimensional imaging system and method of the present invention, a three-dimensional imaging system using a binocular camera and a projector will be introduced first.

[0045] like Figure 1 As shown, the binocular 3D imaging system includes a left camera, a stripe projector, and a right camera. The stripe projector projects orthogonal phase-shifting fringes onto the target object. The left and right cameras simultaneously capture images of the target object with orthogonal phase-shifting fringes. During the calibration phase, the target object serves as a standard image target. During 3D reconstruction, the target object becomes the object to be reconstructed in 3D.

[0046] The three-dimensional coordinates of any point on the target object are: The image coordinates of this point in the corresponding points of the left and right camera images are respectively and In homogeneous coordinates, the linear model, neglecting lens distortion, is expressed as:

[0047] ;

[0048] in:

[0049] The extrinsic parameters for the left camera include rotation and translation parameters;

[0050] This is the internal reference of the left camera;

[0051] This is the internal reference of the right camera;

[0052] These are the focal lengths of the left and right cameras, respectively;

[0053] These are the structural parameters of the binocular system, which are related to the extrinsic parameters calibrated by the left and right cameras.

[0054] It can be any scaling factor;

[0055] , These are the coordinates of the principal points of the left and right cameras, respectively.

[0056] These are the homogeneous image coordinates of the left and right cameras without distortion, respectively.

[0057] The above linear module represents any point in space. The intrinsic and extrinsic parameters of the left and right cameras are converted into points in the camera coordinate space through mathematical operations. .

[0058] After the left and right cameras simultaneously capture images of the target object with orthogonal phase-shifting fringes, phase analysis can be used to obtain orthogonal phase distribution maps from both cameras. , and , For a point in the left camera image Its orthogonal phase value is Find the right camera that satisfies the orthogonal phase value. point . and That is, the corresponding point of the same point on the target object in the left and right camera images.

[0059] During 3D reconstruction, any pixel in the left camera... By using orthogonal phase correspondence, the corresponding pixel on the right camera can be obtained. The corresponding points for distortion removal can be obtained through iteration using the intrinsic parameter matrix and the distortion model formula. , By solving for the principal points of the left and right cameras , Departure process , Corresponding light , The intersections are used for three-dimensional reconstruction.

[0060] Due to system calibration error and corresponding point search error Points to remove distortion , The image planes of the left and right cameras diffuse into an uncertain region. Assume error. and This is a random error, and it follows a normal distribution with a mean of 0 as the U and V of the imaging surface change. Therefore, the uncertain regions are respectively radii of... , A circle, like Figure 1 As shown by the dashed line. That is... In thought center, They appear randomly within a circle of radius [radius value]. Affected by system calibration error; The light beam is affected by both system calibration error and the accuracy of corresponding point search. , They degenerate into two oblique cones respectively. , The corresponding light rays on the left and right sides during the 3D reconstruction process. , The uncertain space at the intersection point consists of two oblique cones, one on the left and one on the right. , Overlapping areas ,like Figure 2 As shown.

[0061] The embodiments of the present invention aim to more accurately determine the spatial coordinates of a point M on a target object in an uncertain space.

[0062] like Figure 3 The diagram shown is a flowchart of a multi-camera-based 3D imaging method according to an embodiment. This method is applied to a 3D imaging system including multiple cameras and a stripe projector. In this embodiment, multiple cameras refer to three or more cameras. The method includes:

[0063] S102: Imaging the target object using a calibrated 3D imaging system. Specifically, the target object projected with orthogonal fringes is imaged using the calibrated 3D imaging system. In the calibrated 3D imaging system, the relative positional relationships between cameras and between each camera and the fringe projector are represented by a set of parameters. Based on these parameters, points on different camera images can be mapped, allowing for 3D reconstruction of the target object from multiple perspectives. In one embodiment, step S102 may further include: calibrating the 3D imaging system to obtain the intrinsic and extrinsic parameters of each camera.

[0064] S104: Using any one of the multiple cameras as the first camera, for the pixels in the image of the target object captured by the first camera, find the sub-pixel points in the images of the target object captured by other cameras that correspond to the pixels in the image of the first camera. Specifically, the multiple cameras acquire orthogonal fringes to obtain an orthogonal phase distribution map for each camera, and use an orthogonal phase analysis method to analyze the orthogonal phase distribution map to obtain the sub-pixel points in the images captured by other cameras that correspond to the pixels in the image of the first camera.

[0065] S106: For the pixels in the image of the first camera and the sub-pixels in the images of other cameras, normalization and distortion correction are performed to obtain the projection points on the normalized imaging plane; among them, the projection points corresponding to the same point on the target object are grouped together. In the camera's local coordinate system, the normalized imaging plane is the plane with camera coordinate Z=1. In practice, camera lenses inevitably have distortion, and third-order radial distortion and second-order tangential distortion are usually considered.

[0066] Specifically, the normalization is performed on distorted pixels on the camera according to the following formula. The distortion-normalized coordinates are obtained by projecting the image onto the normalized imaging plane. :

[0067] ;

[0068] Where f is the focal length of the camera. The coordinates are the principal point of the camera. The distorted pixels include pixels in the image captured by the first camera and sub-pixels in the images captured by other cameras.

[0069] The distortion correction is performed by normalizing the distortion coordinates according to the following formula. Convert to distortion-free normalized coordinates :

[0070] ;

[0071] in, , The radial distortion coefficient is... denoted as the tangential distortion coefficient.

[0072] S108: For each set of projection points on the normalized imaging plane, draw multiple straight lines from the principal point of each camera pointing to the corresponding camera projection point; among them, the straight lines drawn corresponding to the same set of projection points constitute a set. A set of projection points refers to the projection points on the normalized imaging plane corresponding to the same point on the target object and the corresponding points on each camera image after normalization and distortion correction.

[0073] S110: For each set of lines, obtain the coordinates of the spatial point in space that minimizes the sum of the Euclidean distances from each line in the set, and use them as the reconstructed 3D spatial coordinates of the corresponding point on the target object.

[0074] like Figure 4 As shown, in one embodiment, obtaining the coordinates of the spatial point that minimizes the sum of Euclidean distances from each of the set of lines includes:

[0075] S202: Obtain the expression for each line in the coordinate system of any j-th camera; where, N is the number of cameras. In one embodiment, the expression for each straight line in the coordinate system of the j-th camera is represented in point normal form as follows:

[0076] ;

[0077] in Represents the coordinates of a point on the j-th straight line. Let represent the normal vector of the j-th line corresponding to the i-th pixel. represents the coordinates of the principal point of the j-th camera; t represents the proportional coefficient of the distance between a point on a straight line in space and the principal point of the camera.

[0078] S204: Establish spatial points The expression for the distance to the j-th line The j-th straight line corresponds to the j-th camera. In one embodiment, the spatial point... The expression for the distance to the j-th line is:

[0079] ;

[0080] in:

[0081] .

[0082] S206: Establish an expression for the sum of the Euclidean distances of each line in this set of lines:

[0083] ;

[0084] S208: Regarding the above The minimum value is calculated to obtain the coordinates of the required spatial point.

[0085] The aforementioned 3D imaging method employs multiple cameras (three or more). Based on the principle of the aforementioned binocular camera system, it can be known that the coordinates of point M on the target object lie within an uncertain region where three or more conical regions intersect. Figure 5As shown. Therefore, the range of the uncertain area is smaller, allowing for more precise positioning and improving the accuracy of 3D reconstruction.

[0086] The following section details the processing steps for 3D imaging using N cameras.

[0087] Orthogonal fringe projections are acquired using N cameras, and orthogonal phase distributions can be obtained for each camera using phase analysis. Let's define one of the N cameras as camera 1, and consider the image from camera 1. All pixels on the image, assuming one of them is a pixel. Its pixel position is The orthogonal phase value corresponding to camera 1 Using the orthogonal phase value of camera 1, a bilinear search is used to find the camera whose phase values ​​in two orthogonal directions are equal to... subpixel position , , ..., …

[0088] After obtaining the corresponding points, based on the lens distortion model, an iterative distortion correction method is used to obtain the image projection points on the normalized imaging plane. , , ,…, In the camera's local coordinate system, the normalized imaging plane is camera coordinate plane 1, and the three-dimensional coordinates of this projection point are as follows: , , , ..., Starting from the principal point of N cameras , , , ..., Passing through , , , ..., Multiple light rays from a point are arbitrary on one camera. The corresponding light ray. In N different coordinate systems, N light rays can describe the path passing through the principal point. , , , ..., The normal vector is , , , ..., According to the point normal form, the equation of the line in space is:

[0089] ;

[0090] Due to limitations in system accuracy and noise, the intersection of N light rays in reality is not an ideal spatial point. This paper proposes a method using Newton-Raphson iterative optimization to find the spatial point with the shortest sum of distances to N lines. Firstly, the system parameters... , , ..., Transform the N line equations to the camera coordinate system. Its normal vector in the camera coordinate system is... Represented as:

[0091] ;

[0092] The principal point of N cameras In camera coordinate system The Chinese character is represented as:

[0093] ;

[0094] Normals of N straight lines and the main point The point-normal form of the line equation for camera 1 in the same coordinate system is:

[0095] ;

[0096] a point in space The straight-line distance to camera number 1 is:

[0097] ;

[0098] in, ;

[0099] From the above The expression can be similarly used to derive this point. Distance to other cameras , … Therefore, the objective function for optimization is: This is a typical nonlinear optimization problem. Therefore, the Jacobian determinant and the Hessian matrix are used to represent the first and second derivatives, respectively, and the optimal solution is obtained by using the Newton-Raphson iterative method.

[0100] The residual of each line is represented using vector form, where the i-th line originates from the camera's optical center. and unit direction vector Given that the vector residual from point Q to the i-th line is represented by the cross product:

[0101] ;

[0102] in, ,and,

[0103] ;

[0104] Define the antisymmetric matrix of the direction vector:

[0105] ;

[0106] but ,remember:

[0107] ;

[0108] but:

[0109] ;

[0110] The objective function is:

[0111] ;

[0112] Expanding the objective function yields a quadratic form:

[0113] ;

[0114] in, .

[0115] The first derivative with respect to Q is:

[0116] ;

[0117] Hessian (second derivative matrix):

[0118] ;

[0119] The identity of an antisymmetric matrix is:

[0120] ;

[0121] Where I is the identity matrix.

[0122] at the same time: ;

[0123] Therefore, we can conclude that:

[0124] , ;

[0125] Solve using the Newton-Raphson iterative method:

[0126] Hessian is always equal to 2S, therefore:

[0127] ;

[0128] Iterative convergence criterion:

[0129] ;

[0130] After the iterations converge, the optimal solution is obtained. .

[0131] Traversing Camera 1 image By examining all pixels on the graph, we can obtain the coordinates of all pixels in the effective region.

[0132] The embodiments of the present invention use a four-camera system for experimental verification.

[0133] The experiment employed four Hikvision MV-CU013-80GM area array industrial cameras (resolution 1280*1024; pixel size: 4.0μm*4.0μm; frame rate: 90fps) and an 8mm focal length industrial lens (HIK RoBOT) to form a four-camera imaging system. The structured light generator was a custom model from Anhua Optoelectronics (M110B-LC) with a resolution of 1024*768. The system mainly consisted of four cameras (top, bottom, left, and right) and a stripe projector located in the center.

[0134] First, calibration was performed using a 17×15 black background with white dots, a center diameter of 4mm, and a center-to-center spacing of 18mm. The calibration target size was 300mm×200mm, slightly larger than the measurement field of view. During system calibration, the target size was slightly larger than the system's effective field of view. The target underwent 13 different orientations, and the reference points uniformly filled the entire measurement space. At each orientation, four cameras acquired images of the target. The four-camera system was then calibrated based on the 13 different target images acquired by each camera.

[0135] To practically verify the reconstruction accuracy, a standard plane was used as the benchmark, and the plane fitting error of the reconstruction results was used to characterize the 3D reconstruction accuracy. The standard plane was calibrated by a professional institution, and its flatness error was 5μm. The measured results show that the measurement accuracy based on four-camera reconstruction is significantly higher than that based on two-camera reconstruction. Figure 6A and Figure 6B The images shown are pseudo-color images of 3D reconstructions of a planar target using a four-camera and a two-camera system, respectively. The color variations represent the differences from the standard plane. It can be seen that... Figure 6A The color change is less than Figure 6B The color changes indicate higher precision.

[0136] In one embodiment, a three-dimensional imaging system is provided, comprising a plurality of cameras, a stripe projector, a memory, and a processor; the plurality of cameras capture images of a target object onto which orthogonal stripes are projected by the stripe projector; the memory stores a computer program; and the processor executes the computer program to process the image of the target object, thereby implementing the steps of the method described above.

[0137] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the above method steps.

[0138] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the above-described method steps.

[0139] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided by this invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided by this invention may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided by this invention may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0140] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0141] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A three-dimensional imaging method based on multiple cameras, characterized in that, The method, applied to a three-dimensional imaging system comprising multiple cameras and a stripe projector, includes: The target object is photographed using a calibrated 3D imaging system; Using any one of the multiple cameras as the first camera, find the sub-pixel points in the images of the target object captured by the first camera that correspond to the pixels in the images of the target object captured by other cameras. For the pixels in the image of the first camera and the sub-pixels in the images of other cameras, normalization and distortion correction are performed to obtain the projection points on the normalized imaging plane; among them, the projection points corresponding to the same point on the target object are a group. For each set of projection points on the normalized imaging plane, multiple straight lines are drawn from the principal point of each camera, pointing to the corresponding camera projection point; among them, the straight lines drawn from the same set of projection points constitute a set. For each set of straight lines, obtain the coordinates of the spatial point that minimizes the sum of Euclidean distances from each line in the set, and use these coordinates as the reconstructed 3D spatial coordinates of the corresponding point on the target object; obtaining the coordinates of the spatial point that minimizes the sum of Euclidean distances from each line in the set includes: Obtain the expression for each line in the coordinate system of any j-th camera; where, N is the number of cameras; Establish spatial points The expression for the distance to the j-th line The j-th straight line corresponds to the j-th camera; Establish an expression for the sum of the Euclidean distances of each line in this set of lines: ； Regarding the The minimum value is calculated to obtain the coordinates of the desired spatial point. The expression for each straight line in the coordinate system of the j-th camera, expressed in point-normal form, is as follows: ； in Represents the coordinates of a point on the j-th straight line. Let represent the normal vector of the j-th line corresponding to the i-th pixel. represents the coordinates of the principal point of the j-th camera; t represents the proportionality coefficient of the distance between a point on a straight line in space and the principal point of the camera. The expression for the distance from a point in space to each straight line is: ； in: 。 2. The multi-camera-based three-dimensional imaging method according to claim 1, characterized in that, Also includes: The three-dimensional imaging system is calibrated to obtain the intrinsic and extrinsic parameters of each camera.

3. The multi-camera-based three-dimensional imaging method according to claim 1, characterized in that, The normalization is performed according to the following formula to normalize the distorted pixels on the camera. The distortion-normalized coordinates are obtained by projecting the image onto the normalized imaging plane. : Where f is the focal length of the camera. The coordinates are the principal point of the camera, and the distorted pixels include pixels on the image captured by the first camera and sub-pixels on the images captured by other cameras.

4. The multi-camera-based three-dimensional imaging method according to claim 3, characterized in that, The distortion correction is performed by normalizing the distortion coordinates according to the following formula. Convert to distortion-free normalized coordinates : ； in, , The radial distortion coefficient is... denoted as the tangential distortion coefficient.

5. The multi-camera-based three-dimensional imaging method according to claim 1, characterized in that, The step of using a calibrated 3D imaging system to photograph a target object includes: using a calibrated 3D imaging system to photograph a target object on which orthogonal stripes are projected by the stripe projector; The step of finding sub-pixel points on images of the target object captured by other cameras that correspond to pixels on the image of the first camera includes: the multiple cameras acquiring orthogonal fringes to obtain an orthogonal phase distribution map for each camera, and using an orthogonal phase analysis method to analyze the orthogonal phase distribution map to obtain sub-pixel points on images captured by other cameras that correspond to pixels on the image of the first camera.

6. The multi-camera-based three-dimensional imaging method according to claim 1, characterized in that, The solution is obtained using the Newton-Raphson iterative method.

7. A multi-camera-based three-dimensional imaging system, comprising multiple cameras, a stripe projector, a memory, and a processor; wherein the multiple cameras capture images of a target object onto which orthogonal stripes are projected by the stripe projector, and the memory stores a computer program, characterized in that... When the processor executes the computer program to process the image of the target object, it implements the steps of the method according to any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images