13 November 2014 A new technique of recognition for coded targets in optical 3D measurement
Author Affiliations +
Abstract
A new technique for coded targets recognition in optical 3D-measurement application is proposed in this paper. Traditionally, point cloud registration is based on homologous features, such as the curvature, which is time-consuming and not reliable. For this, we paste some coded targets onto the surface of the object to be measured to improve the optimum target location and accurate correspondence among multi-source images. Circular coded targets are used, and an algorithm to automatically detecting them is proposed. This algorithm extracts targets with intensive bimodal histogram features from complex background, and filters noise according to their size, shape and intensity. In addition, the coded targets’ identification is conducted out by their ring codes. We affine them around the circle inversely, set foreground and background respectively as 1 and 0 to constitute a binary number, and finally shift one bit every time to calculate a decimal one of the binary number to determine a minimum decimal number as its code. In this 3Dmeasurement application, we build a mutual relationship between different viewpoints containing three or more coded targets with different codes. Experiments show that it is of efficiency to obtain global surface data of an object to be measured and is robust to the projection angles and noise.
© (2014) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Changye Guo, Xiaosheng Cheng, Haihua Cui, Ning Dai, Jinping Weng, "A new technique of recognition for coded targets in optical 3D measurement", Proc. SPIE 9276, Optical Metrology and Inspection for Industrial Applications III, 92761H (13 November 2014); doi: 10.1117/12.2070957; https://doi.org/10.1117/12.2070957
PROCEEDINGS
7 PAGES


SHARE
Back to Top