Detecting, identifying, quantifying and tracking strategies for leakage plastic litter in terrestrial as well as aquatic environments using remote sensing technologies is expected to support future continuous monitoring of waste pollution at wide covering geo-spatial scales. The methodologies to detect plastic waste from airborne or satellite observations should be based on fundamental and well-characterised optical features. Among the important optical characteristics is the polarisation of light by plastics which we presume could be useful to better understand remote sensing information in addition to reflectance shape and magnitude. Spectropolarimetry is expected to contribute towards approaches that mitigate uncertainties resulting from sea surface reflected glint and atmospheric correction (e.g., Gilerson et al., 2020; Harmel and Chami, 2013).

Polarimetric properties of virgin and weathered plastic litter are not well documented in open-access (Garaba and Harmel, 2022), therefore in the framework of the ESA Discovery Campaign on Remote Sensing for Plastic Marine Litter we proposed to conduct relevant groundwork. The research aimed at developing algorithms for data handling, processing and documenting polarisation characteristics of the tested plastics. In addition, a theoretical study was conducted to propagate the plastic optical signature within the atmosphere-water system with a specific focus on intensity and polarisation of light at the top-of-atmosphere.

A set of three sensor technologies were evaluated with the main goal to establish a reference library of polarimetric properties of plastics. Imaging was conducted using a SPECIM-IQ hyperspectral camera covering the visible to near infrared (VNIR, 40-1000 nm) spectrum whilst additional data was captured using a compact multispectral Teledyne FLIR BlackFly S USB3 BFS-U3-51S5PC-C polarimetric true colour Red-Green-Blue (RGB) sensor. Point measurements (~400-900 nm) were also collected using the unique hyperspectral spectropolarimeter GroundSPEX (van Harten et al., 2014). Open-source Python-based packages were developed and adapted to analyse data from the above sensors. We believe the developed open-source packages are simple to use, easy to adapt and powerful tools for scientific efforts to handle acquired images and provide quality-controlled datasets of plastic litter polarisation properties that can also be easily adapted for other studies. Regarding the theoretical work, two different data sets were produced: (i) top-of-atmosphere signature of floating and submerged macroplastics, (ii) in-water, above-water and TOA polarization signature (Stokes vector) of nano and micro-plastic particles in suspension.

Project Team: Dr. Shungu Garaba (University of Oldenburg, Germany), Dr. Tristan Harmel (TH-Consulting, France), Dr. Olivier Burggraaff (Leiden University, Netherlands).

The prime's project webpage can be found here.