ESAPlastics – De-Risk Action: Spectrometer for Marine Litter


This De-Risk Action will evaluate and develop at a low TRL, prospective technology that in the future can lead to the development of a spectrometer for marine litter detection from space. Clearly there is an opportunity to develop innovative technology that can take advantage of Earth Observation satellite constellations to provide a continuous monitoring of earth oceans and coastal areas.

Marine litter is considered a global problem, and if no change occurs in the near future it will be even bigger by the end of 2030. Therefore this De-Risk activity will follow a two-step technical approach:

  1. Acquisition of reliable marine litter data, using heterogeneous sensors (visible cameras, multispectral cameras, hyperspectral cameras in the VNIR and SWIR range) located at different altitudes, and using different systems, namely: satellite PRISMA and Sentinel 2 information, Unmanned Aerial Vehicles (UAV) and manned aircraft.
  2. Processing of such information, leading to two valuable outcomes; one is machine learning methods to help classify marine litter using artificial intelligence (AI) in post processing; the other, the theoretical assessment and evaluation of compressive sensing techniques using tailored spectral bands, in view of its potential improved SNR, minimizing mass, volume and/or power, as a potential technique for high sensitivity satellite based plastic monitoring.

This project is partially funded by the European Space Agency’s (ESA) under the General Support Technology Programme (GSTP) to test sensor-based perception technologies capable of detecting plastics in coastal areas and shallow waters.


INESC Technology and Science – INESC TEC is an Associate Laboratory with more than 30 years of experience in R&D and technology transfer. It is a private non-profit research institution having as associates the University of Porto, INESC and the Polytechnic Institute of Porto. With around 800 researchers (350 PhD), works in the interface between the academic world and the industrial and service companies, as well as the public administration. The activity at INESC TEC runs under the paradigm of the knowledge to value production chain: knowledge and results generated at basic research are typically injected in technology transfer projects and therefore they receive added social relevance. The existence of an Innovation and Technology Transfer Unit assures the effectiveness of this model. INESC TEC incorporates 13 R&D Centres and one Associate Unit with complementary competences, always looking to the international market.

The main contributions from INESC TEC to this project will be made by the Centre for Robotics and Autonomous Systems (CRAS) and the Centre for Applied Photonics (CAP).CRAS conducts research and development activities in autonomous robotic systems, mobile robotics and mobile multi-robot systems for inspection, monitoring and mapping, with applications in security, power systems, environment, aquaculture, oceanography, marine biology, resource extraction, among other sectors. These activities are supported by the research in perception, navigation, control, localization, coordination, and automatic data collection and processing. CAP detains long experience with lasers, spectroscopy and optical sensors.

The AIR Centre is an internationally networked organization, oriented to foster job creation and knowledge-driven sustainable economic development in Atlantic regions. It addresses and integrates space, climate, earth, ocean, energy and data sciences and promotes north-south/south-north/north-north/south-south cooperation in alignment with national/regional priorities and global challenges such as the UN 2030 Agenda for Sustainable Development, the Decade of Ocean Science for Sustainable Development (2021-2030), the Paris Agreement and the Sendai Framework for Disaster Risk Reduction.

The AIR Centre is all about advancing science and technology in a transformative scale in the Atlantic region. It builds on and expands the abilities of individual organizations, and it advances selected scientific and technological domains and their constellations of actors towards shared targets. For that, it recruits and orchestrates a complex web of organizations and individuals to deliver change and social impact through concrete actions.

This complexity comes from the AIR Centre’s unique multidimensional mission-oriented, demand-driven, problem-solving approach, which integrates various sciences (space, ocean, earth, climate, energy, and data sciences), includes different stakeholders (government, academia, industry, and civil society), encompasses diverse geographies, cultures and technology readiness levels (American, African, European countries and small island states or territories in the Atlantic region), and fully accommodates both local priorities and global challenges.


The Institute of Marine Research (IMAR) was created with a mission to “conduct leading deepsea and open ocean research and education to advance the understanding of marine systems in a changing planet, and promote the sustainable blue economy and management of marine ecosystems for the benefit of the society and the environment”. IMAR develops both fundamental and applied research with many international partners from around the world, and promotes strong links with governmental and private organizations, and with the society as a whole.

IMAR is a leading national research centre on marine sciences recently affiliated to the MARE Marine and Environmental Sciences Centre. IMAR currently includes 31 PhD Integrated Members, 54 regular members (technicians, junior researchers, and students), and various collaborating members. IMAR activities are directed to scientific research and education on the ecology, ecosystem functioning, and ecosystem based management of open ocean and deep-water marine ecosystems, including seamounts, hydrothermal vents, mid-ocean ridges, abyssal plains, and the pelagic and mesopelagic realms, with a special focus on the Atlantic Ocean but with extensions to other ocean basins.

IMAR has contributed profoundly, and will keep contributing to the sustainable management of the oceans at the local, national, european, and international levels though a dedicated agenda which promotes the transfer of scientific knowledge to governmental and private organizations, and to the society. IMAR has also paid special attention to improving ocean literacy and environmental awareness.


In this WP, all management and reporting issues of the project will be dealt. This WP will ensure a correct and on-time execution of the tasks according with the predicted. Moreover, in this workpackage some dissemination will be performed according to the results of the de-risk action.

This WPs will be devoted to an update of the state of the art of available technologies to perform spectral detection of plastics. It will have as an input recent results from ESA ongoing projects, and will dedicate some attention to new emerging technologies such as compressive sensing techniques in the context of marine litter detection, taking into consideration a potential future adaption to satellite-borne detection.

This task will define the type of campaigns needed according to the requirements of the systems to be implemented. Airborne campaigns will be set (with UAV and aircrafts) covering different ranges in altitude and different systems in complexity and payload. Concerning the hyperspectral imaging sensor, the objective is to put a hyperspectral imaging sensor onboard an Unmanned Aerial Vehicle (UAV). With the hyperspectral sensor on board we will detect, identify and classify Marine Litter based on its spectral signature. To accomplish this objective, datasets will be acquired in near coastal areas, rivers and beaches. Next, the algorithms that allow to distinguish marine litter from the cluttered background will be fine-tuned. Based on this procedure and after the required algorithm training, INESC TEC developed already processing tools will be potentially used to detect the presence of marine litter using hyperspectral imaging data.

This WP will involve three tasks:

Task 4.1 – This task will run in parallel and close collaboration with Task 4.3. From the spectral data and classification algorithms, the most relevant spectral bands, its importance and characteristics (width, spectral weight) in plastic litter automatic detection will be identified. This information will lead to the prospection of commercial available technologies (detectors and optical hardware) that can be used to produce a compressive sensing hyperspectral laboratory system. The optical performance will be theoretically assessed by proper simulations. The hardware implementation will be tested, adapting the compressive sensing platform already developed in the CYCLOPS/ESA and performance parameters will be experimentally assessed as a validation for the performed simulations.

Task 4.2  – Advanced Spectral characterization of the collected samples. The objective of this task is to evaluate different established spectral characterization techniques as tools to identify and characterize microplastics and associated contaminants. A database of spectra acquired with different techniques from established contaminants will be built and used as reference. Collected samples will be characterized and compared against the database enabling the implementation of compositional maps showing the distribution of the different plastics in the samples. Techniques to be tested in an exploratory basis can include UV-VIS-NIR, LIBs and Raman spectroscopy.

Task 4.3 – Post-processing for Plastic Detection based on multispectral/hyperspectral image classification (airborne and satellite image comparison). The objective of this task is to study and develop image classification methods using spatial and spectral image information based on post-processing approaches that allow to classify marine litter (i.e. plastics) using available airborne data e.g UAV, manned airplanes and satellite data. To achieve this is necessary to develop the processing algorithms using the acquired data in task 4’.2 and to test the different algorithms using remotely acquired data (satellite) or manned airplane campaigns.

Additional Information

Image Gallery

Video Gallery