Sub-pixel detection method for an aviation plug based on Zernike moments
Ji Li, Si Long Qin, Gang Han, Wei Zhao, Huai Jiang Yin · Applied Optics · 2025
To address the multi-aperture detection requirements for aviation plugs in complex aerospace manufacturing environments, a sub-pixel edge detection method based on Zernike moments was proposed. First, a linear ramp edge detection model is constructed on the basis of traditional Zernike moments. By precisely deriving edge parameters in the fuzzy transition region via orthogonal critical moments and integrating an adaptive extraction algorithm to optimize threshold judgment criteria, the sub-pixel localization accuracy is significantly improved. Second, a local optimization mechanism and dynamic parameter adjustment strategy were introduced to refine the RANSAC fitting algorithm. Finally, a high-precision visual inspection system for aviation plugs was developed, integrating the Zernike moment-based sub-pixel edge detection algorithm. The experimental results show that the absolute error between the proposed method and the coordinate measuring machine measurements is ±10µm, with a root mean square error of 2.99 µm and a detection time of 1.016 s, demonstrating high accuracy and excellent real-time detection performance in complex industrial environments.