Towel mark recognition based on HOG for cloth handling robot system

Author

Fukumoto, Yasuhiko ; Hamada, Toshihiro ; Kobayashi, Hiroaki ; Hata, Seiji ; Hayashi, Kenta ; Hojoh, Hirotaka

Author_Institution

Kagawa Prefectural Ind. Technol. Center, Kagawa, Japan

fYear

2010

fDate

7-10 Nov. 2010

Firstpage

1198

Lastpage

1202

Abstract

To automate linen supply factories completely, we are developing a flexible objects handling robot system. The purpose of this system is to input washing clothes to the pressing and folding line which is still operated by human operators. Now its targets are rectangular clothes as a first step. In order to handle towels, the system has to recognize which client´s towel is it and which side up is it. Therefore, we studied about towel mark recognition, and adopted the feature descriptors based on Histograms of Oriented Gradients (HOG). The results of experiment show the success rate is almost 100%.

Keywords

factory automation; feature extraction; flexible manipulators; industrial manipulators; laundering; materials handling; robot vision; HOG; cloth handling robot system; histograms of oriented gradient; linen supply factory; towel mark recognition; Correlation; Grasping; Histograms; Production facilities; Robot sensing systems;

fLanguage

English

Publisher

ieee

Conference_Titel

IECON 2010 - 36th Annual Conference on IEEE Industrial Electronics Society

Conference_Location

Glendale, AZ

ISSN

1553-572X

Print_ISBN

978-1-4244-5225-5

Electronic_ISBN

1553-572X

Type

conf

DOI

10.1109/IECON.2010.5675539

Filename

5675539