Investigation and optimization of a novel precision measurement method based on template matching

Abstract

Polar microstructure is the 10-μm aperiodic surface pattern which is designed to achieve nanoscale resolution and applied in the precision measurement area. In this paper, an investigation of the major factors of a novel template-matching based precision measurement method which integrates with polar microstructure and computer vision (IPMCV) is proposed. In order to optimize the measurement performance of IPMCV method, three crucial factors are determined according to the template-matching method: the size of global image, the size of template image and the number of interpolated sub-pixels. According to the measurement accuracy, the size of global image has little influence on the position error, but significantly influences the standard deviation value of the error gradually. Besides, the size of template image helps to significantly improve the measurement accuracy, and the increase of the number of interpolated sub-pixels increases the measurement accuracy within a certain range but has no effect afterwards. The experimental results have great significance on the determination and optimization of parameters for IPMCV method to achieve a better measurement performance.

A panoramic image stitching method for nuclear fuel assemblies

“…The panoramic image stitching methods have been widely studied over the past few decades [10][11][12][13][14][15]. Image registration is the core of image stitching [16], aiming to determine the geometric relationship between multiple images [17].…”