We have been supporting MSc research into the use of high resolution imagery and multi-spectral sensing for the automated identification and mapping of coastal plastic pollution. Plastic pollution within the marine environment is a huge global challenge. The scale of the problem is continuously increasing and consequently there is an urgent need to refine methods to detect and quantify plastic debris in marine and coastal environments to facilitate mitigation strategies. Over recent years we have supported numerous MSc research projects to investigate methods for remotely detecting and quantifying plastics within the coastal and nearshore environment. Typically we work at a remote beach site at Tyne Sands, East Lothian, where we can conduct controlled experiments with different types of plastics, sensors and mission parameters. Across the various projects we have assessed various drone-based cameras and specialised multispectral and hyperspectral imagers (which use only specific wavebands to identify different materials) along with object and pixel based image analysis and machine learning techniques. To produce a truly useful operational tool, sensors would need to be flown on a manned aircraft than can cover an entire coastline in an hour or two, so among the key questions investigated are the spatial and spectral resolutions required to achieve successful automated detection, and hence the implications for sensor specification and mission parameters required for a manned aircraft survey. For this work we have typically deployed our Matrice 300 RTK (now with P1 45 Mp camera), Mavic 3 Multispectral drone, Mavic 2 Pro, and historically our T680 drone with Parrot Sequoia multispectral camera. For future work we can also deploy the Maia S2 multispectral camera, which has bands aligned to the Sentinel-2 satellite system. This article was published on 2024-06-25
