Effective Structural Health Monitoring of Rotating Propellers using Asynchronous Neuromorphic Tracking 127 Fig. 4 Detected propeller at different timestamps Fig. 5 Angles in radians calculated by the code (a) (b) Fig. 6 Displacement tests conducted. (a) Laser directly pointing at propeller; (b) Laser reflected through galvanometer Regarding the integrated operation of all system components—namely, the detection of the propeller, angle calculation, information transmission, and mirror position updating—tests were performed where the propeller was spun at varying speeds to observe the mirror movements. The results, shown in Figure 7, indicate that the system can accurately detect the propeller, process information, calculate the angle, and adjust the mirrors accordingly. However, it is important to note that the Python code is not yet optimized, and current hardware limitations create processing delays, causing the laser to move at the same speed as the propeller with a slight lag. Data acquisition and processing Initially, the code utilized time intervals to group data, process it, and iterate through the cycle. This approach led to varying data group sizes based on rotational speed, causing slow responses at high speeds and significant delays. To address this issue, the decision was made to implement event-based grouping, maintaining a quasi-constant iteration time regardless of rotational speed. Furthermore, the original approach involved using two separate Python scripts: one optimized for controlling the camera and processing all data, and the other for operating the data acquisition system (DAQ) and generating the analog signal. It became apparent that communication issues between the scripts increased delays significantly, prompting the integration of both functionalities into a single multifunctional script.
RkJQdWJsaXNoZXIy MTMzNzEzMQ==