Automated Non-Contact Vibration Measurement with Dynamic Vision: Framework and Signal Component Analysis

Vibration analysis is of great importance for condition monitoring of mechanical systems. Non-contact vibration measurement methods are generally more flexible for application. However, deployment calibration is usually complex and difficult in practice. To address such issues, we propose an automated non-contact vibration measurement framework with dynamic vision. By utilizing the event stream filtering method, a vibration signal reconstruction algorithm is proposed, which is implemented with a robot arm through feedback control. A quantitative evaluation method is established for evaluating the efficiency of vibration component analysis. Compared with the traditional contact vibration sensors through experimental validation, the proposed method demonstrates significant potential in analyzing machine vibrations and performing fault diagnosis. Through experimental validation on the pump test rig, it has been validated that the proposed framework is highly efficient and user-friendly. That shows promising application prospects in the real industrial scenarios.