Evaluating the detect, identify, quantify and track advances in remote sensing of aquatic plastic litter

• Applicant: Dr. Shungudzemwoyo Garaba
• Subject Area: Geodesy, Photogrammetry, Remote Sensing, Geoinformatics, Cartography
• Term: since 2018
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 417276871

Leakage plastics (LPs) have become a wicked environmental problem for mankind especially the blue economy. As the problem has grown, interdisciplinary evidence-based studies have also increased, yet experts still suggest little is known about LPs given the vastness of the oceans. The data on counts and distributions of LPs are still temporally and geographically sparse. These data gaps are typically filled in and inferred from sparse net trawl counts after regressing them against numerical model solutions that tend to exhibit large uncertainties. One strategy that has gained interest as a solution to filling in the gaps has been integrating remote sensing (RS) and net trawl counts into numerical modelling. Already, direct and indirect RS observations of LPs are being leveraged and explored in a rising number of experiment-based studies. These works have been showcasing and assessing the potential of RS to ‘detect’ presence and classify plastics in an imagery, ‘identify’ polymer type, ‘quantify’ individual items, patch sizes or pixel coverage and ‘track’ pathways, source, sinks. The proposed work here builds off the groundwork by Dr. Garaba in the visible to shortwave infrared spectrum by further studying hyperspectral imagery, thermal and polarized spectral properties of LPs. This proposed work will be based on the FAIR policy to foster open-access to datasets and will aim to make algorithms open-source. To this end, the expected outcomes from this proposed disruptive ‘open’ science include (i) establishing and expanding imagery and spectral reference libraries obtained using prototype and standard RS technologies, (ii) evaluating artificial intelligence through deep learning algorithms to generate essential LPs descriptors, (iii) extensive two-way transfer of knowledge through stakeholder engagement within the recently formed IOCCG Task Force on Remote Sensing of Marine Litter and Debris and (iv) benchmarking of current as well as emerging technologies and algorithms relevant to remote sensing of plastic litter. These expected outcomes are considered challenging but feasible as preliminary investigations have shown promise, thus knowledge gained can be leveraged and harnessed by state-of-the-art RS technologies generating high geo-spatial, repeated and hyperspectral imagery for the purposes of detecting, identifying, quantifying and tracking aquatic plastics. A continuous reassessment of algorithms and technologies will be expected as technology is rapidly evolving in an interdisciplinary field that is considered in its infancy. Ideally, RS should support stakeholders with affordable and sustainable evidence-based data streams to mitigate LPs, guide clean-up actions and thus grow a circular economy.

• DFG Programme: Research Grants