Mapping coastal marine debris: Application for UAV remote sensing and object-based image analysis

Marine debris from anthropogenic sources is a global issue that has shown to produce adverse effects on marine wildlife and ecosystems, health, and ecosystem services and economies. Despite the severity of marine debris, of which plastic debris constitute 60 – 80 %, uncertainties exist in the estimation of marine mass balance. In order to develop effective strategies for marine debris control, remediation, and prevention, informed decision-making surrounding debris mass balance and standing stock is critical. Recently, remote sensing has been put forward as a potential solution to current bottlenecks that exist in traditional debris monitoring in marine environments, however manual analysis of remote sensing images remains a slow process. This study tests object-based image analysis (OBIA) as an automated method of monitoring debris in remote sensing images. The study also compares the effectiveness of unmanned aerial vehicles (UAVs) against other remote sensing platforms. The study found that the high variation of debris, coupled with the high variation of coastal environment background, created challenges in identifying debris through OBIA. Missing debris and false positives were considerable in sites with a high degree of heterogeneity of debris and background. The study also found that UAVs are an effective platform for monitoring large-scale regions with orthomosaics and that slight changes in cloud cover can complicate image detection. Ultimately the study highlighted remote sensing and automated image analysis challenges that still exist for coastal marine debris monitoring