High-Speed 3D Motion Attitude Measurement Technology Based on a Monocular Vision Multi-camera Array

In response to the limitations in usage scenarios and low camera utilization of 3D pose measurement methods based on binocular vision or multiple camera intersections for high-speed moving objects, a 3D motion pose measurement method using a monocular vision high-speed camera array is adopted. This paper proposes a monocular visual pose measurement algorithm for the object under test based on the <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$PnP$</tex> (Perspective-n-Point) algorithm. By constructing a likelihood function, the camera distortion coefficients are solved and optimized. Utilizing the one-to-one correspondence between the theoretical and actual coordinates of the object under test, the complete 3D motion pose of the object is obtained through an orthogonal iterative solution algorithm and a stitching algorithm. The mathematical models involved in this method and their derivation processes are introduced, and experiments are conducted to verify the effectiveness of the method.