View Item 

Feature extraction of non-uniform food products using RGB and RGB-D data combined with shape models

Access Status
Open access
Authors
Wurdermann, H.
Aminzadeha, V.
Cui, Lei
Dai, J.
Date
2011
Type
Conference Paper

Metadata
Show full item record
Citation
Wurdermann, Helge A. and Aminzadeh, Vahid and Cui, Lei and Dai, Jian S. 2011. Feature extraction of non-uniform food products using RGB and RGB-D data combined with shape models, in Wang, Z. (ed), 2011 IEEE International Conference on Robotics and Biomimetics (ROBIO), Dec 7-11 2012, pp. 1652-1657. Phuket, Thailand: IEEE Press.
Source Title
Proceedings of the 2011 IEEE International Conference on Robotics and Biomimetics (ROBIO)
Source Conference
2011 IEEE International Conference on Robotics and Biomimetics (ROBIO)
Additional URLs
ISBN
Remarks

© 2011 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

URI
Collection
Abstract

This research paper investigates a new 3D handling process in food industry: bin picking. Machines only function effectively when the input of product is physically well organised, well structured, and consistent. At many stages in a typical production line, foodstuffs are physically arranged as they move through a machine or equipment, however, this order is then lost again as products are ejected onto conveyors, bulked together into bins for transport, taken off-line for chilled storage. Bin picking is generally not solved for manufacturing parts. Unlike food ordering processes such as pick and place operations, vibratory feeders etc., this food handling operation has not been applied to food industry neither. A new approach is presented using the Microsoft KinectTM sensor and Active Shape Models. By combining the new device that obtains an RGB and RGB-D image and the flexible shape model, it is possible to identify non-uniform food products that have a high variation in shape. The methodology of this system is presented. The experiments show the achievability of this new system.