Adaptive Subpixel Edge Detection for Locating the Center of Nut Screw Hole
- Resource Type
- Article
- Authors
- Rui Yang; Hanming Guo; Zhentao Chen; Jiuai Sun
- Source
- International Journal of Precision Engineering and Manufacturing, 22(8), pp.1357-1364 Aug, 2021
- Subject
- 기계공학
- Language
- English
- ISSN
- 2234-7593
An accurate and fast machine vision-based method for locating the center of disk nuts is of tremendous significant for safe installation and maintaining service life of glass curtain wall. However, current manual or image processing approaches suffer from the problems of inaccurate or time consuming. Based on the intensity distribution model around the edge of screw holes within actual captured images, a Gaussian weighted adaptive threshold method is proposed to replace traditional Otsu threshold algorithm to identify the edge of screw holes in a level of subpixel. The identified edge points are used to fi t the center of hole via a least square estimation algorithm. Both simulation and real object evaluation have shown that the proposed algorithm has higher accuracy in locating the center of the screw holes and demonstrated with good tolerance to noise and faster processing speed comparing to that of the traditional algorithm.