Integrated Vision Tools and Signature Analysis for Part Measurement and Recognition in Robotic Tasks
| Authors: | Borangiu Theodor, University Politehnica of Bucharest, Faculty of Control and Computers, Romania
Anton Florin, University Politehnica of Bucharest, Faculty of Control and Computers, Romania
Tunaru Silvia, University Politehnica of Bucharest, Faculty of Control and Computers, Romania
Ivanescu Andrei-Nick, University Politehnica of Bucharest, Faculty of Control and Computers, Romania |
|---|
| Topic: | 4.3 Robotics |
|---|
| Session: | Robot Control II |
|---|
| Keywords: | robot vision, object recognition, signature analysis, inspection, learning |
|---|
Abstract
In this paper, aspects concerning image analysis and interpretation for signature analysis are presented. The first part of this paper analyses the most useful sets of region- and boundary features for shape description; such descriptors allow mapping the object space into the feature space which will be used in statistical pattern recognition for cluster building and pattern identification. Standard intrinsic features are predefined global functions applicable to any class of objects; they are extended with local and global measurements which are applied to particular classes of objects. Measurements are based on vision tools, like windows region of interest (WROI), linear and arc rulers, and point-, line- and arc finders; they can be integrated in complex signature measurements for shape analysis. The paper is devoted to object measurement and recognition by signature analysis.
