Paper Abstract and Keywords |
Presentation |
2022-11-24 15:20
Locating the Fruit to Be Harvested and Estimating their Cut Positions from RGBD Images Acquired by a Camera Moved along a Fixed Path Using a Mask-RCNN Based Method Wentao Zhao, Jun Ohya, Chanjin Seo, Takuya Otani, Taiga Tanaka, Koki Masaya, Atsuo Takanishi (Waseda Univ.), Shuntaro Aotake, Masatoshi Funabashi (SONY CSL) CS2022-55 IE2022-43 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
This paper proposes a Mask R-CNN[1] based method for locating fruits (tomatoes and yellow bell peppers, etc.) and estimating the cutting positions from RGBD images acquired by a camera moved along a fixed path. After getting mask results of all fruits and pedicels (cutting positions), the proposed method judges the matching relationship between the located fruit and pedicel according to the distance between fruit and pedicel. Experimental results show the proposed method effectively detects the cutting position of each fruit. The method is also robust in complex environments. In addition, it turns out that the fixed path strategy is valid for avoiding obstacles and reaching the pedicel and cutting position accurately. A high harvesting success rate was achieved in a Gazebo based simulated environment. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Harvest Robot / Point Cloud Processing / Mask R-CNN / Locating Cutting position / Fixed Harvesting path / / / |
Reference Info. |
IEICE Tech. Rep., vol. 122, no. 270, IE2022-43, pp. 39-44, Nov. 2022. |
Paper # |
IE2022-43 |
Date of Issue |
2022-11-17 (CS, IE) |
ISSN |
Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
CS2022-55 IE2022-43 |
|