Search

CN-116630434-B - Binocular structured light imaging method and system based on lifting platform calibration

CN116630434BCN 116630434 BCN116630434 BCN 116630434BCN-116630434-B

Abstract

The invention relates to the technical field of machine vision, and discloses a binocular structured light imaging method and a binocular structured light imaging system based on lifting platform calibration, which can obtain imaging effects with high precision and good stability by providing a new model for calculating 3D coordinates and solving the model by adopting an iterative method; meanwhile, matching points are reversely calculated according to the monocular imaging result and the calibration model, so that accuracy of searching the matching points is improved, whole lines do not need to be searched each time, and time consumption of searching is reduced.

Inventors

  • CHEN SHUN
  • KOU GUANZHONG
  • PAN WEI
  • CAO LING
  • LU SHENGLIN

Assignees

  • 广东奥普特科技股份有限公司

Dates

Publication Date
20260505
Application Date
20230428

Claims (12)

  1. 1. A binocular structured light imaging method based on lift table calibration, the method comprising: s1, calibrating a binocular camera and a projector respectively; s2, searching matching points of the left camera and the right camera by constructing the same world coordinates, and constructing 17 parameter models for all the matching points to calculate binocular 3D coordinates; s3, solving model parameters by adopting a Gaussian Dessel iteration method to obtain X, Y, Z dual-target fixed-conversion matrixes in three directions 、 、 ; S4, solving an absolute phase diagram for the calibration plate fringe diagram, and solving corresponding projector pixel coordinates; s5, solving according to a standard pinhole camera model to obtain a left camera homography matrix and a right camera homography matrix 、 And projector homography matrix 、 ; S6, obtaining left and right camera absolute phase diagrams for imaging objects 、 ; S7, simultaneous And Solving a corresponding 3D coordinate of the left camera, and solving a 3D coordinate of the right camera in the same way; s8, according to the 3D coordinate sum of the left camera Solving for pixel coordinates on the corresponding right camera, and based on the right camera 3D coordinates and Solving the pixel coordinates on the corresponding left camera; S9, according to the pixel coordinates and the absolute phase diagram corresponding to the 3D coordinates obtained in the step S8 、 Searching left and right camera matching points in a fixed interval; S10, determining a conversion matrix according to the matching points and the double targets 、 、 And obtaining binocular 3D coordinates.
  2. 2. The binocular structured light imaging method based on lift table calibration of claim 1, wherein step S1 comprises: a calibration plate is selected and placed on the lifting platform; the lifting table is controlled to lift at different heights, and calibration plate images at different heights are collected; Under different heights, collecting more than 6 groups of calibration plate images to calibrate the binocular camera; and under different heights, calibrating the projector by acquiring calibration slat images with the X direction and the Y direction being greater than 6X 24.
  3. 3. The binocular structured light imaging method based on lift table calibration of claim 2, wherein step S2 comprises: Finding matching points of left and right cameras by constructing the same world coordinates And ; The following 17-parameter model was constructed for all matching points for computing binocular 3D coordinates: ; ; 。
  4. 4. A binocular structured light imaging method based on lift calibration according to claim 3, wherein step S3 comprises: Solving model parameters by using Gaussiedel iteration method to obtain X, Y, Z dual-target fixed-conversion matrix in three directions 、 、 The iterative formula is as follows: 。
  5. 5. The binocular structured light imaging method of claim 4, wherein step S4 comprises: Defining the absolute phase diagram of X, Y directions as 、 Image points The width and height of the projector are pw and ph, and the pixel coordinates of the projector are obtained The calculation formula is as follows: ; ; ; The same is done to obtain: 。
  6. 6. the binocular structured light imaging method of claim 5, wherein step S5 comprises: Solving according to a standard pinhole camera model to obtain a left camera homography matrix and a right camera homography matrix 、 And projector homography matrix 、 The formula is as follows: ; The format rewritten as ax=b is as follows: ; obtaining coefficients by direct solution 。
  7. 7. The binocular structured light imaging method of claim 6, wherein step S7 comprises: Combined stand And The corresponding 3D coordinates of the left camera are obtained, and the 3D coordinates of the right camera are obtained by the same method, and the formula is as follows: ; 、 、 ; 、 、 ; 、 、 ; 、 、 ; ; ; ; ; ; ; ; ; 。
  8. 8. the binocular structured light imaging method of claim 7, wherein the step S9 comprises: the pixel coordinate corresponding to the 3D coordinate is (u, v), defining 、 Traversing Points in (a) When meeting the following requirements And performing linear interpolation processing, wherein the formula is as follows: ; ; 。
  9. 9. The binocular structured light imaging method of claim 8, wherein step S10 comprises: matching points obtained in the previous step And Is carried into the formula in the step S2, and the conversion matrix is determined according to the double targets 、 、 And obtaining binocular 3D coordinates.
  10. 10. A binocular structured light imaging system based on lift table calibration, the system comprising: The calibration module is used for calibrating the binocular camera and the projector respectively; The model construction module is used for searching matching points of the left camera and the right camera by constructing the same world coordinates, and constructing 17 parameter models for all the matching points to calculate binocular 3D coordinates; the model iteration module is used for solving model parameters by adopting a Gaussian Dessel iteration method to obtain X, Y, Z dual-target fixed-conversion matrixes in three directions 、 、 ; The first pixel coordinate solving module is used for solving an absolute phase diagram for the calibration plate fringe diagram and solving corresponding projector pixel coordinates; the matrix solving module is used for solving and obtaining left and right camera homography matrixes according to the standard pinhole camera model 、 And projector homography matrix 、 ; A phase diagram solving module for solving the absolute phase diagrams of the left and right cameras for the imaging object 、 ; Camera 3D coordinate solving module for simultaneous use And Solving a corresponding 3D coordinate of the left camera, and solving a 3D coordinate of the right camera in the same way; a second pixel coordinate solving module for solving the three-dimensional coordinate sum according to the 3D coordinate sum of the left camera Solving for pixel coordinates on the corresponding right camera, and based on the right camera 3D coordinates and Solving the pixel coordinates on the corresponding left camera; A matching point searching module for searching the pixel coordinates and the absolute phase map corresponding to the 3D coordinates obtained in the step S8 、 Searching left and right camera matching points in a fixed interval; The binocular 3D coordinate solving module is used for S10, determining a conversion matrix according to the matching points and the double targets 、 、 And obtaining binocular 3D coordinates.
  11. 11. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor, when executing the computer program, implements a binocular structured light imaging method based on lift calibration as claimed in any one of claims 1-9.
  12. 12. A storage medium containing computer executable instructions for execution by a computer processor to implement the lift-table calibration based binocular structured light imaging method of any one of claims 1-9.

Description

Binocular structured light imaging method and system based on lifting platform calibration Technical Field The invention relates to the technical field of machine vision, in particular to a binocular structured light imaging method and system based on lifting platform calibration. Background Machine vision is a branch of the rapid development of artificial intelligence. In short, machine vision is to simulate the visual function of a person by using a computer, extract information from an image of an objective object, process and understand the information, and finally use the information for actual detection, measurement and control. The binocular imaging technology is one of the machine vision technologies, in most binocular imaging systems, three-dimensional images of objects can be generated through the steps of three-dimensional calibration, three-dimensional correction, three-dimensional matching, 3D coordinate calculation and the like, and because the three-dimensional matching adopts a matching algorithm to search matching points of left and right cameras, the whole process is very time-consuming, the calculated 3D coordinate model is relatively fixed, and the image accuracy is more dependent on the matching effect. Compared with the binocular imaging technology, the binocular structured light imaging technology has higher efficiency, lower data processing capacity and wider application. However, even this, the binocular structured light imaging technique still has the problems of low calibration accuracy, unstable accuracy, too long time for searching left and right matching points, and the like. Accordingly, improvements in the art are needed. The above information is presented as background information only to aid in the understanding of the present disclosure and is not intended or admitted to be prior art relative to the present disclosure. Disclosure of Invention The invention provides a binocular structured light imaging method and a binocular structured light imaging system based on lifting platform calibration, which are used for improving the calibration precision, precision stability and imaging efficiency of a binocular structured light technology. In order to achieve the above object, the present invention provides the following technical solutions: In a first aspect, the present invention provides a binocular structured light imaging method based on lift calibration, the method comprising: s1, calibrating a binocular camera and a projector respectively; s2, searching matching points of the left camera and the right camera by constructing the same world coordinates, and constructing 17 parameter models for all the matching points to calculate binocular 3D coordinates; S3, solving model parameters by adopting a Gaussian Dessel iteration method to obtain a X, Y, Z double-target fixed-conversion matrix H x、Hy、Hz in three directions; s4, solving an absolute phase diagram for the calibration plate fringe diagram, and solving corresponding projector pixel coordinates; S5, solving according to a standard pinhole camera model to obtain a left camera homography matrix H cl、Hcr and a right camera homography matrix H pl、Hpr; s6, obtaining left and right camera absolute phase maps I l、Ir of an imaging object; S7, simultaneously H cl and H pl calculate corresponding 3D coordinates of the left camera, and similarly calculate 3D coordinates of the right camera; S8, resolving the pixel coordinate on the corresponding right camera according to the 3D coordinate of the left camera and H cr, and resolving the pixel coordinate on the corresponding left camera according to the 3D coordinate of the right camera and H cl; S9, searching left and right camera matching points in a fixed interval according to pixel coordinates corresponding to the 3D coordinates obtained in the step S8 and an absolute phase diagram I l、Ir; S10, obtaining binocular 3D coordinates according to the matching points and the double-target fixed transformation matrix H x、Hy、Hz. In one possible implementation manner, in the binocular structured light imaging method based on lift table calibration, step S1 includes: a calibration plate is selected and placed on the lifting platform; the lifting table is controlled to lift at different heights, and calibration plate images at different heights are collected; Under different heights, collecting more than 6 groups of calibration plate images to calibrate the binocular camera; and under different heights, calibrating the projector by acquiring calibration slat images with the X direction and the Y direction being greater than 6X 24. In one possible implementation of the method according to the invention, in the binocular structured light imaging method based on the calibration of the lifting platform, the step S2 comprises the following steps: Finding matching points (L u,Lv) and (R u,Rv) of the left and right cameras by constructing the same world coordinates; The following 17-parameter mode