Object Recognition and Tracking in Moving Videos for Maritime Autonomous Surface Ships

被引：19
作者：
Park, Hyunjin [1 ]
Ham, Seung-Ho [2 ]
Kim, Taekyeong [1 ]
An, Donghyeok [1 ]
机构：
[1] Changwon Natl Univ, Dept Comp Engn, Chang Won 51140, South Korea
[2] Changwon Natl Univ, Dept Naval Architecture & Marine Engn, Chang Won 51140, South Korea
来源：
JOURNAL OF MARINE SCIENCE AND ENGINEERING | 2022年 / 10卷 / 07期
基金：
新加坡国家研究基金会;
关键词：
object recognition; object tracking; deep learning; maritime autonomous surface ship;
D O I：
10.3390/jmse10070841
中图分类号：
U6 [水路运输]; P75 [海洋工程];
学科分类号：
0814 ; 081505 ; 0824 ; 082401 ;
摘要：
In autonomous driving technologies, a camera is necessary for establishing a path and detecting an object. Object recognition based on images from several cameras is required to detect impediments in autonomous ships. Furthermore, in order to avoid ship collisions, it is important to follow the movements of recognized ships. In this paper, we use the Singapore Maritime Dataset (SMD) and crawling image for model training. Then, we present four YOLO-based object recognition models and evaluate their performance in the maritime environment. Then, we propose a tracking algorithm to track the identified objects. Specially, in evaluation with high-motion video, the proposed tracking algorithm outperforms deep simple online and real-time tracking (DeepSORT) in terms of object tracking accuracy.
引用
收藏
页数：17
相关论文
共 24 条
  • [21] Wojke N, 2017, IEEE IMAGE PROC, P3645, DOI 10.1109/ICIP.2017.8296962
  • [22] Encoding Crowd Interaction with Deep Neural Network for Pedestrian Trajectory Prediction
    Xu, Yanyu
    Piao, Zhixin
    Gao, Shenghua
    [J]. 2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2018, : 5275 - 5284
  • [23] Pedestrian trajectory prediction with convolutional neural networks
    Zamboni, Simone
    Kefato, Zekarias Tilahun
    Girdzijauskas, Sarunas
    Noren, Christoffer
    Dal Col, Laura
    [J]. PATTERN RECOGNITION, 2022, 121
  • [24] Zhang Y., 2020, ARXIV, DOI [10.1007/s11263-021-01513-4, DOI 10.1007/S11263-021-01513-4]
123