Nebula Public Library

The “Detection of Ocean Litter Plastics with Hyper-to-multispectral Infrared Neural Networks” (DOLPHINN) study explored a novel image processing technique to improve the use of space assets to map the distribution of plastic across the globe. The DOLPHINN project objective was to determine the feasibility of using a hyperspectral-to-multispectral association machine learning approach to improve the capability of space-based short wave infrared (SWIR) multispectral assets to detect, quantify, and track plastic litter in waters and on shorelines.

For decades humans have been producing plastic materials and this production has never ceased growing to a point that, due to waste mismanagement, a global environmental issue has been generated. One of the most mediatic manifestations of such a pollution is the Great Pacific Garbage Patch (GPGP), an area in the North Pacific ocean with higher density of macro and microplastic representing between three and six times the size of France .

The main aim of OPTIMAL was to perform a feasibility study on a potential mission or system to detect marine plastic litter through satellite remote sensing. To do so, the project activities involved the following aspects:

The depletion of stratospheric ozone due to anthropogenic emissions of long-lived substances containing chlorine and other halogens was first identified as a matter of international concern in the mid-1970s and has received increasing attention since the mid-1980s after the discovery of the ozone hole. Today, in spite of the successful application of the Montreal Protocol and its amendments, the atmospheric halogen loading is still above natural levels and will remain so until the middle of this century.