POSTER SESSION 1
15:30–17:30, Monday, September 14
Poster Session | 1 | 2 | 3 | 4 | Schedule at a Glance
ABSTRACT 1411 | POSTER 109
FIVE-YEAR MONITORING OF COASTAL MARINE DEBRIS ON UNINHABITED ISLANDS USING HIGH-RESOLUTION OPTICAL SATELLITE IMAGERY
Coastal marine debris causes environmental pollution, damages tourism resources, and marine ecosystems. However, regular field surveys are difficult on remote uninhabited islands because these areas are hard to access. Optical satellite imagery provides an effective way to observe large coastal areas repeatedly and consistently.
In this study, we used high-resolution optical satellite images with a spatial resolution of 0.5 m or better to detect coastal debris on uninhabited islands and analyze changes over a five-year period. To improve the consistency and reliability of the results, all images were processed using atmospheric correction, land and cloud masking, and orthorectification. Coastal debris was detected using an ensemble machine-learning model that combined Support Vector Machine (SVM), Random Forest, and XGBoost.
The model achieved an accuracy of 99.35%, a precision of 77.42%, a recall of 78.00%, and an F1 score of 77.71%, showing good performance for detecting coastal debris on uninhabited islands. Monthly analysis of debris distribution identified periods of debris accumulation and increase on individual islands and revealed seasonal patterns over the five-year study period.
These results show that high-resolution optical satellite imagery can be used to effectively monitor the spatial and temporal distribution of coastal debris and support decision-making for debris management and cleanup activities in remote island environments.
Taeho Kim, Underwater Survey and Technology 21 Inc., [email protected], https://orcid.org/0000-0002-7664-1539
Ho-Kun Jeon, Underwater Survey and Technology 21 Inc., [email protected]
Poster Session | 1 | 2 | 3 | 4 | Schedule at a Glance
