Instance segmentation models for detecting floating macroplastic debris from river surface images

被引:0
作者
Kataoka, Tomoya [1 ,2 ]
Yoshida, Takushi [3 ]
Yamamoto, Natsuki [3 ]
机构
[1] Ehime Univ, Dept Civil & Environm Engn, Matsuyama, Japan
[2] Ehime Univ, Ctr Marine Environm Studies, Matsuyama, Japan
[3] Yachiyo Engn Co Ltd, Business Planning & Dev Div, Tokyo, Japan
关键词
floating macroplastic debris; transport; YOLOv8; instance segmentation; river surface; fixed camera; ultrasonic water level gauge;
D O I
10.3389/feart.2024.1427132
中图分类号
P [天文学、地球科学];
学科分类号
07 ;
摘要
Quantifying the transport of floating macroplastic debris (FMPD) in waterways is essential for understanding the plastic emission from land. However, no robust tool has been developed to monitor FMPD. Here, to detect FMPD on river surfaces, we developed five instance segmentation models based on state-of-the-art You Only Look Once (YOLOv8) architecture using 7,356 training images collected via fixed-camera monitoring of seven rivers. Our models could detect FMPD using object detection and image segmentation approaches with accuracies similar to those of the pretrained YOLOv8 model. Our model performances were tested using 3,802 images generated from 107 frames obtained by a novel camera system embedded in an ultrasonic water level gauge (WLGCAM) installed in three rivers. Interestingly, the model with intermediate weight parameters most accurately detected FMPD, whereas the model with the most parameters exhibited poor performance due to overfitting. Additionally, we assessed the dependence of the detection performance on the ground sampling distance (GSD) and found that a smaller GSD for image segmentation approach and larger GSD for object detection approach are capable of accurately detecting FMPD. Based on the results from our study, more appropriate category selections need to be determined to improve the model performance and reduce the number of false positives. Our study can aid in the development of guidelines for monitoring FMPD and the establishment of an algorithm for quantifying the transport of FMPD.
引用
收藏
页数:14
相关论文
共 24 条
[1]  
Ahmed D, 2023, Arxiv, DOI arXiv:2304.13282
[2]  
Al-Zawaidah H, 2021, ENVIRON SCI-PROC IMP, V23, P535, DOI [10.1039/d0em00517g, 10.1039/D0EM00517G]
[3]   CM-YOLOv8: Lightweight YOLO for Coal Mine Fully Mechanized Mining Face [J].
Fan, Yingbo ;
Mao, Shanjun ;
Li, Mei ;
Wu, Zheng ;
Kang, Jitong .
SENSORS, 2024, 24 (06)
[4]   Diverging estimates of river plastic input to the ocean [J].
Gonzalez-Fernandez, Daniel ;
Roebroek, Caspar T. J. ;
Laufkoetter, Charlotte ;
Cozar, Andres ;
van Emmerik, Tim H. M. .
NATURE REVIEWS EARTH & ENVIRONMENT, 2023, 4 (07) :424-426
[5]  
Hao Y., 2022, arXiv
[6]   Learning non-maximum suppression [J].
Hosang, Jan ;
Benenson, Rodrigo ;
Schiele, Bernt .
30TH IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2017), 2017, :6469-6477
[7]   Measuring riverine macroplastic: Methods, harmonisation, and quality control [J].
Hurley, Rachel ;
Braaten, Hans Fredrik Veiteberg ;
Nizzetto, Luca ;
Steindal, Eirik Hovland ;
Lin, Yan ;
Clayer, Francois ;
van Emmerik, Tim ;
Buenaventura, Nina Tuscano ;
Eidsvoll, David Petersen ;
Okelsrud, Asle ;
Norling, Magnus ;
Adam, Hans Nicolai ;
Olsen, Marianne .
WATER RESEARCH, 2023, 235
[8]   Detecting the interaction between microparticles and biomass in biological wastewater treatment process with Deep Learning method [J].
Jia, Tianlong ;
Peng, Zhaoxu ;
Yu, Jing ;
Piaggio, Antonella L. ;
Zhang, Shuo ;
Kreuk, Merle K. de .
SCIENCE OF THE TOTAL ENVIRONMENT, 2024, 951
[9]   Advancing deep learning-based detection of floating litter using a novel open dataset [J].
Jia, Tianlong ;
Vallendar, Andre Jehan ;
de Vries, Rinze ;
Kapelan, Zoran ;
Taormina, Riccardo .
FRONTIERS IN WATER, 2023, 5
[10]   Deep learning for detecting macroplastic litter in water bodies: A review [J].
Jia, Tianlong ;
Kapelan, Zoran ;
de Vries, Rinze ;
Vriend, Paul ;
Peereboom, Eric Copius ;
Okkerman, Imke ;
Taormina, Riccardo .
WATER RESEARCH, 2023, 231