Vibration Mode Identification Method of Large-scale Flexible Structure Based on Computer Vision

Aiming at the measurement and identification of vibration modes of large flexible attachments on orbiting spacecraft, this paper proposes a method for modal parameter identification of large flexible structures based on computer vision. This paper expounds the related technologies of image processing, the imaging model of monocular camera, the extraction algorithm of cooperative target centroid and the measurement data processing algorithm. A kind of ground satellite simulation experiment system with a lightweight and large flexible truss structure is tested. The monocular camera is used to collect and extract the star coordinates of the cooperative target in real time. The coordinates are unified according to the motion model and the Kalman filter is used to eliminate noise interference. The modal vibration frequency information of the measurement point is obtained by analysis. The analysis of identification results shows that this method can identify the low-order vibration frequency error of flexible structures less than 5%, which has high engineering practice value.