POLYPROBLEM report: The Circularity Code

Remote sensing of plastic emissions FROM A BIRD’S EYE VIEW Although digital tools make documentation much easier, removing waste from the environment is tedious and usually done by hand; often selectively and above all, it requires committed individuals. Considering the enormous scale of the plastics crisis, complementary approaches are needed to remove pollution that has already occurred. Remote sensing is playing an increasingly important role in this field. Surveillance cameras, aircraft, drones, and satellites record and examine the physical properties of an area from a distance by measuring reflected and emitted radiation.73 AI as your right hand PlasticObs_plus is a project that has set out to use aircraft to detect floating and washed-up plastic waste from the air in order to draw conclusions about its origins and distribution routes. “In the future, we want to Source: Blume et al. (2023),p. 19 73 see Blume et al. (2023), p. 11 Recording, analyzing, and deriving knowledge from remote sensing images Collecting images by remote sensing ─ Satellites used (e.g., Sentinel-2, WorldView-3, GeoEye-1, PlanetSope, SkySat, PRISMA, EnMAP) ─ Aerial platforms (e.g., drones, aircraft) ─ Land-based platforms (e.g., boats, fixed posts, bridges, handheld devices) Validation and verification ─ Monitoring/observation on site ─ Known or artificial targets ─ Social media ─ Citizen science Preprocessing images ─ Atmosphere correction ─ Land mask creation ─ Cloud detection ─ Detection of white foam ridges ─ Sun glint correction ─ Labeling of known objects Analysis and classification ─ Pixel coverage analysis ─ Spectral properties (shape and brightness) ─ Algorithms (e.g., band index, normalized differences, band depth, neural networks, object edge detection, random forest) ─ Sensitivity tests Output ─ Detection, quantification and identification of end products (e.g., maps, quantities, polymers, shapes, colors, sizes) ─ Statistical metrics (e.g., precision, pixel coverage, false-positive and false-negative results, F-score) 48

RkJQdWJsaXNoZXIy ODI5MzU=