LAUSR

To evaluate the assembly accuracy of rudders during the production of aerospace vehicles, a measurement method based on monocular vision is proposed in this paper. Compared with existing methods that adopt cooperative targets pasted manually, the proposed method avoids pasting cooperative targets on the surface of rudders and calibrating the original position of rudders in advance. First, we use two known position markers on the surface of the vehicle and multiple feature points of the rudder to solve the relative pose between the camera and the rudder by employing the PnP algorithm. Then, we measure the rotation angle by converting the change of the camera's pose to the rotation angle of the rudder. Finally, a tailored error compensation model is introduced into the proposed method to increase the accuracy of the measurement. Experiment results show that the average measurement absolute error of the proposed method is less than 0.08° overall, which is remarkably superior to existing methods and satisfies the requirement of practical industrial production.