High-efficiency sub-microscale uncertainty measurement method using pattern recognition

Chenyang Zhao,Chi Fai Cheung,Peng Xu
DOI: https://doi.org/10.1016/j.isatra.2020.01.038
IF: 7.3
2020-06-01
ISA Transactions
Abstract:<p>This study presents a fast precision measurement method that uses pattern recognition. First, a specific micro-structured surface was designed and manufactured, providing a unique pattern for recognition and matching. Second, a measurement system was proposed based on the algorithms of circle Hough transform (CHT), neural classifier (NC), template matching (TM) and sub-pixel interpolation (SI). Then, a series of experiments were carried out from three aspects: circle detection, length uncertainty, and measurement speed and range. The results showed the correct circle classification percentage was more than 96% and the CHT search accuracy was within a two-pixel level. The length uncertainty test demonstrated the method was able to achieve 90-nm length uncertainty, and a comparison of measurement speeds showed it helped to speed up measurements by a factor of 1000 compared to the original one.</p>
automation & control systems,instruments & instrumentation,engineering, multidisciplinary
What problem does this paper attempt to address?