Research on Dimensional Measurement Based on Sub-pixel Edge Detection

Abstract

With the development of modern industry, image measurement technology with its high speed ,high precision and non-contact advantages receives high-profile attention. In the machine vision system the size of the process of mechanical parts measured, it is found that the accuracy of the edge position directly influence the accuracy of the measurement results. According to the research on the current sub-pixels positioning of image processing technology, this paper firstly makes a theory analysis and research about sub-pixel location methods based on gray level moment theory and the theory of Gaussian fitting. Then through the parts size measurement experiment, under the premise of contrast location with classical edge detection operators, some groups of data are extracted respectively compared to both of the detection performance and accuracy.Thus it provided the reference for the sub-pixels edge detection algorithm in the actual application.