Tan Chi Jie1, Takumi Tomokawa1, Shintaro Ogawa1, Ayumu Tominaga2, Sakmongkon Chumkamon1, Eiji Hayashi1
1Department of Mechanical Information Science and Technology, Kyushu Institute
of Technology, 680-4, Kawazu, Iizuka-City, Fukuoka, 820-8502, Japan
2Department of Creative Engineering Robotics and Mechatronics Course, National
Institute of Technology Kitakyushu College, 5-20-1 Shii, Kokuraminamiku,
Kitakyushu, Fukuoka, 802-0985, Japan
pp. 224–229
ABSTRACT
Datasets are one of the key elements which determine the performance of
a deep learning network. Urban environments datasets receive much attention
nowadays due to the rise of autonomous cars but off-road environment on
the other hand lacks quality datasets. Offroad environments need equal
attention as only 55% of the world’s population lives in urban areas. This
paper tackles this issue to close the gap of robotic visual perception
on the beach, one of the common offroad environments that lack attention
by presenting a real and synthetic dataset, namely BCRobo.
ARTICLE INFO
Article History
Received 20 November 2022
Accepted 28 June 2023
Keywords
Synthetic dataset
Computer vision
Real dataset
Field robotics
Image segmentation
JAALR3406
Download article(PDF)