Accuracy improvement of semantic segmentation using appropriate datasets for robot navigation

Ryusuke Miyamoto, Miho Adachi, Yuta Nakamura, Takeshi Nakajima, Hiroki Ishida, Shingo Kobayashi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

The use of detailed metric maps for autonomous movement of robots has been popularized in recent times. Three-dimensional sensing devices, such as 3D LiDAR and RADAR, which are expensive yet indispensable, are utilized to generate these metric maps, and ultimately, perform localization. To reduce the cost of sensing devices, we try to realize autonomous movement of a robot using only cheap image sensors, such as webcams. For robot navigation, image processing tends to be applied to collision avoidance by finding obstacles. In contrast, our approach does not use visual object detection but adopts path planning based on movable area extraction from input images using semantic segmentation. To obtain accurate results for visual navigation, this paper proposes the use of novel datasets for semantic segmentation. Experimental results showed that ICNet could extract the movable area with more than 99% accuracy if it was trained with appropriate datasets, and a robot can run automatically based on the extracted movable area.

Original languageEnglish
Title of host publication2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1610-1615
Number of pages6
ISBN (Electronic)9781728105215
DOIs
Publication statusPublished - Apr 2019
Event6th International Conference on Control, Decision and Information Technologies, CoDIT 2019 - Paris, France
Duration: 23 Apr 201926 Apr 2019

Publication series

Name2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019

Conference

Conference6th International Conference on Control, Decision and Information Technologies, CoDIT 2019
CountryFrance
CityParis
Period23/04/1926/04/19

Fingerprint

Robot Navigation
Navigation
Segmentation
Robot
Semantics
Robots
Sensing
Metric
Image Sensor
Collision Avoidance
Object Detection
Lidar
Path Planning
Image Processing
Collision avoidance
Motion planning
Image sensors
Tend
Three-dimensional
Image processing

Cite this

Miyamoto, R., Adachi, M., Nakamura, Y., Nakajima, T., Ishida, H., & Kobayashi, S. (2019). Accuracy improvement of semantic segmentation using appropriate datasets for robot navigation. In 2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019 (pp. 1610-1615). [8820616] (2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CoDIT.2019.8820616
Miyamoto, Ryusuke ; Adachi, Miho ; Nakamura, Yuta ; Nakajima, Takeshi ; Ishida, Hiroki ; Kobayashi, Shingo. / Accuracy improvement of semantic segmentation using appropriate datasets for robot navigation. 2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 1610-1615 (2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019).
@inproceedings{084a5ff8884f49faa941628269da11ad,
title = "Accuracy improvement of semantic segmentation using appropriate datasets for robot navigation",
abstract = "The use of detailed metric maps for autonomous movement of robots has been popularized in recent times. Three-dimensional sensing devices, such as 3D LiDAR and RADAR, which are expensive yet indispensable, are utilized to generate these metric maps, and ultimately, perform localization. To reduce the cost of sensing devices, we try to realize autonomous movement of a robot using only cheap image sensors, such as webcams. For robot navigation, image processing tends to be applied to collision avoidance by finding obstacles. In contrast, our approach does not use visual object detection but adopts path planning based on movable area extraction from input images using semantic segmentation. To obtain accurate results for visual navigation, this paper proposes the use of novel datasets for semantic segmentation. Experimental results showed that ICNet could extract the movable area with more than 99{\%} accuracy if it was trained with appropriate datasets, and a robot can run automatically based on the extracted movable area.",
author = "Ryusuke Miyamoto and Miho Adachi and Yuta Nakamura and Takeshi Nakajima and Hiroki Ishida and Shingo Kobayashi",
year = "2019",
month = "4",
doi = "10.1109/CoDIT.2019.8820616",
language = "English",
series = "2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "1610--1615",
booktitle = "2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019",

}

Miyamoto, R, Adachi, M, Nakamura, Y, Nakajima, T, Ishida, H & Kobayashi, S 2019, Accuracy improvement of semantic segmentation using appropriate datasets for robot navigation. in 2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019., 8820616, 2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019, Institute of Electrical and Electronics Engineers Inc., pp. 1610-1615, 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019, Paris, France, 23/04/19. https://doi.org/10.1109/CoDIT.2019.8820616

Accuracy improvement of semantic segmentation using appropriate datasets for robot navigation. / Miyamoto, Ryusuke; Adachi, Miho; Nakamura, Yuta; Nakajima, Takeshi; Ishida, Hiroki; Kobayashi, Shingo.

2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019. Institute of Electrical and Electronics Engineers Inc., 2019. p. 1610-1615 8820616 (2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Accuracy improvement of semantic segmentation using appropriate datasets for robot navigation

AU - Miyamoto, Ryusuke

AU - Adachi, Miho

AU - Nakamura, Yuta

AU - Nakajima, Takeshi

AU - Ishida, Hiroki

AU - Kobayashi, Shingo

PY - 2019/4

Y1 - 2019/4

N2 - The use of detailed metric maps for autonomous movement of robots has been popularized in recent times. Three-dimensional sensing devices, such as 3D LiDAR and RADAR, which are expensive yet indispensable, are utilized to generate these metric maps, and ultimately, perform localization. To reduce the cost of sensing devices, we try to realize autonomous movement of a robot using only cheap image sensors, such as webcams. For robot navigation, image processing tends to be applied to collision avoidance by finding obstacles. In contrast, our approach does not use visual object detection but adopts path planning based on movable area extraction from input images using semantic segmentation. To obtain accurate results for visual navigation, this paper proposes the use of novel datasets for semantic segmentation. Experimental results showed that ICNet could extract the movable area with more than 99% accuracy if it was trained with appropriate datasets, and a robot can run automatically based on the extracted movable area.

AB - The use of detailed metric maps for autonomous movement of robots has been popularized in recent times. Three-dimensional sensing devices, such as 3D LiDAR and RADAR, which are expensive yet indispensable, are utilized to generate these metric maps, and ultimately, perform localization. To reduce the cost of sensing devices, we try to realize autonomous movement of a robot using only cheap image sensors, such as webcams. For robot navigation, image processing tends to be applied to collision avoidance by finding obstacles. In contrast, our approach does not use visual object detection but adopts path planning based on movable area extraction from input images using semantic segmentation. To obtain accurate results for visual navigation, this paper proposes the use of novel datasets for semantic segmentation. Experimental results showed that ICNet could extract the movable area with more than 99% accuracy if it was trained with appropriate datasets, and a robot can run automatically based on the extracted movable area.

UR - http://www.scopus.com/inward/record.url?scp=85072824113&partnerID=8YFLogxK

U2 - 10.1109/CoDIT.2019.8820616

DO - 10.1109/CoDIT.2019.8820616

M3 - Conference contribution

AN - SCOPUS:85072824113

T3 - 2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019

SP - 1610

EP - 1615

BT - 2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Miyamoto R, Adachi M, Nakamura Y, Nakajima T, Ishida H, Kobayashi S. Accuracy improvement of semantic segmentation using appropriate datasets for robot navigation. In 2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019. Institute of Electrical and Electronics Engineers Inc. 2019. p. 1610-1615. 8820616. (2019 6th International Conference on Control, Decision and Information Technologies, CoDIT 2019). https://doi.org/10.1109/CoDIT.2019.8820616