A Preliminary Analysis towards Detecting Floating Marine Macro Plastics Using an Index Developed for Sentinel 2 ACOLITE and Sen2Cor Images

Abstract

The improper management of plastic waste has become a silent killer in the marine ecosystem. The scientific evidence of plastic accumulated in marine biota and entering the food web have urged quick actions against the plastic pollution. The plastic detection algorithms developed using satellite images can proclaim emission, transportation, weathering and accumulation of plastic that is vital in eliminating them. This study developed an index to find floating marine harvested macro plastic within Sentinel 2 ACOLITE and Sen2Cor images. This index preserved the plastic information protected the plastic information and maximized the separation of them from the surrounding objects. The identified plastic pixels were analysed using scatter plots to discriminate plastic from non-plastic objects and to observe the characteristics. The index and the scatter plot analysis detected the plastic pixels with bottle percentages more than or equal to 14%. The plastic bags and the fishing nets required around 50% pixel percentage to be detected. The plastic pixels with 100% plastic coverage were located as a separate cluster in the scatter plots. Therefore, the accurate detection of plastic pixels using ACOLITE and Sen2Cor images depends on the pixel plastic percentage and plastic coverage. The Sen2cor images classified plastic pixels according to their pixel plastic percentage for two dates. However, this classification was not successful for other dates with aerosol, clouds and smooth sea surface conditions which made the plastic signal weak. The ACOLITE and Sen2Cor atmospheric corrections are not suitable for plastic detection when the plastic signal is weak.