Related Experiment Video
Updated: Jul 30, 2025
10:12
Extraction of Organochlorine Pesticides from Plastic Pellets and Plastic Type Analysis
Published on: July 1, 2017
11.6K
BePLi Dataset v1: Beach Plastic Litter Dataset version 1 for instance segmentation of beach plastic litter
Mitsuko Hidaka1,2, Koshiro Murakami1, Kenta Koshidawa1
1Research Institute for Value-Added-Information Generation (VAiG), Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Kanagawa, Japan.
Data in Brief
|May 14, 2023
Summary
A new dataset of 3,709 images of marine plastic pollution aids the development of automated image analysis tools. This resource supports scientific research and coastal management for cleaner beaches.
Area of Science:
- Environmental Science
- Computer Science
- Data Science
Background:
- Marine plastic pollution is a significant global environmental challenge.
- Automated image analysis is crucial for effective scientific research and coastal management of plastic litter.
Purpose of the Study:
- To introduce the Beach Plastic Litter Dataset version 1 (BePLi Dataset v1).
- To facilitate the development of machine learning models for identifying plastic litter on beaches.
Main Methods:
- The BePLi Dataset v1 contains 3,709 images from diverse coastal environments in Yamagata Prefecture, Japan.
- Images feature various backgrounds like sand, rocks, and tetrapods.
- Instance-based and pixel-level annotations in a modified MS COCO format identify plastic litter (e.g., bottles, fishing gear, foam).
Main Results:
- The dataset provides comprehensive annotations for manual instance segmentation of beach plastic litter.
- It enables the training of machine learning models for precise plastic litter identification.
Conclusions:
- The BePLi Dataset v1 is a valuable resource for advancing automated detection of marine plastic pollution.
- This technology can enhance scalability for estimating plastic litter volume, aiding researchers and government bodies in monitoring and analysis.
