Multi-camera 3D Object Reconstruction for Industrial Automation

In this paper, a method to automate industrial manufacturing processes using an intelligent multi-camera system to assist a robotic arm on a production line is presented. The examined assembly procedure employs a volumetric method for the initial estimation of object’s properties and an octree decomposition process to generate the path plans for the robotic arm. Initially, the object is captured by four cameras and its volumetric representation is produced. Thereafter, a quality check with its respective CAD model is performed and the final details of the 3D model are refined. An octree decomposition technique is utilized afterwards to facilitate the automatic generation of the assembly path plans and translate them to a sequence of movements for the robotic arm. The algorithm is fast, computationally simple and produces an assembly sequence that can be translated to any major robotic programming language. The proposed algorithm is assessed and preliminary experimental results are discussed.

[1]  A. Laurentini,et al.  The Visual Hull Concept for Silhouette-Based Image Understanding , 1994, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  Antti Oulasvirta,et al.  Computer Vision – ECCV 2006 , 2006, Lecture Notes in Computer Science.

[3]  Richard Szeliski,et al.  A Comparison and Evaluation of Multi-View Stereo Reconstruction Algorithms , 2006, 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06).

[4]  Hugh F. Durrant-Whyte,et al.  A Bayesian Approach to Optimal Sensor Placement , 1990, Int. J. Robotics Res..

[5]  Jonathan Corney,et al.  An automated system for the assembly of octree models , 2004 .

[6]  Katsushi Ikeuchi,et al.  Task-oriented generation of visual sensing strategies , 1995, Proceedings of IEEE International Conference on Computer Vision.

[7]  Peter-Pike J. Sloan,et al.  Visual tunnel analysis for visibility prediction and camera planning , 2000, Proceedings IEEE Conference on Computer Vision and Pattern Recognition. CVPR 2000 (Cat. No.PR00662).

[8]  Mark R. Stevens,et al.  Methods for Volumetric Reconstruction of Visual Scenes , 2004, International Journal of Computer Vision.

[9]  Stefano Soatto,et al.  Multi-View Stereo Reconstruction of Dense Shape and Complex Appearance , 2005, International Journal of Computer Vision.

[10]  Gregory D. Hager,et al.  Computational Methods for Task-directed Sensor Data Fusion and Sensor Planning , 1991, Int. J. Robotics Res..

[11]  Anurag Mittal,et al.  Generalized Multi-sensor Planning , 2006, ECCV.

[12]  Theodore Lim,et al.  Automatic generation of robot and manual assembly plans using octrees , 2010 .

[13]  Ruzena Bajcsy,et al.  Occlusions as a Guide for Planning the Next View , 1993, IEEE Trans. Pattern Anal. Mach. Intell..

[14]  Kiriakos N. Kutulakos,et al.  Recovering shape by purposive viewpoint adjustment , 1992, International Journal of Computer Vision.

[15]  Sukhan Lee,et al.  Computer-Aided Mechanical Assembly Planning , 1991 .

[16]  Andrea Cavallaro,et al.  Multi-Camera Networks: Principles and Applications , 2009 .

[17]  Yiming Ye,et al.  Sensor Planning for 3D Object Search, , 1999, Comput. Vis. Image Underst..

[18]  J. O'Rourke.  Art gallery theorems and algorithms , 1987 .

[19]  Roger Y. Tsai,et al.  A versatile camera calibration technique for high-accuracy 3D machine vision metrology using off-the-shelf TV cameras and lenses , 1987, IEEE J. Robotics Autom..