The Adaptive Threshold-Based FFT Spectrum Stripe Edge Detection Method Used for Measuring The High-Frequency Vibration Parameters
Huinan Gong, Ligong Wang, Ming Hai Yang, Qingsong Wu, Deguang Wang, Shengnan Zuo · 2023
On account of its characteristics of non-contact, simplicity, portability, and high accuracy, machine vision has gradually become one of the most flexible and effective methods for measuring vibration parameters. When using machine vision for high-frequency vibration measurement, the captured images inevitably suffer from blurring. To ensure the effectiveness of traditional high-frequency vibration measurement methods based on clear sequential images, image restoration techniques are required to deblurring the blurred images. However, this method has drawbacks such as complex algorithms and low measurement efficiency. This study propose a method for high-frequency vibration parameter measurement using FFT spectrum stripe edge detection on a single blurred image. By combining an improved Canny edge detection algorithm with the Radon transform, we achieve high-precision extraction of the direction and position of stripes in the FFT spectrum of the blurred image, enabling further calculation of the vibration direction and amplitude. The proposed method is evaluated using simulated high-frequency vibration-blurred images with a frequency of 600 Hz. The simulation results demonstrate that the proposed method achieves accurate measurement of high-frequency vibration parameters. In low-amplitude scenarios, the measurement results outperform existing research methods and exhibit good stability.