POSEIDON aims to develop a multi-sensor system mounted on UAV for the detection of floating marine pollution, a serious problem of this society. To this end, POSEIDON proposes a compact system that can be carried out in small UAVs that integrate radar sensors that operate in conjunction with EO/IR cameras to monitor marine areas in order to detect marine debris and oil spills, but that the Time offers an effective detection capacity in the complex conditions of observation of the maritime environment. In particular, the Ku band Polarimetric radar will be used to offer a solution for 24/7, all the meteorological surveillance with greater capacity of detection of floating residues and oil spills, thanks to polarimetry and high spatial resolution. In particular, the use of thermal sensors allows night surveillance, as well as the detection of thermal contrast discharges. For its part, the multispectral image in the VIS/NIR range offers high resolution information that allows the classification of objects and substances on the surface of the sea based on their appearance and spectral signature. The polarimetric process of radar can be used for the improvement of the SCR (disturbance signal-one ratio) to enhance detection performance also in case of adverse sea conditions.
The use of EO/IR sensors is intended to improve the performance of the detection and to provide the identification capacity of the type of pollution and measurement of its extension. To this end, we will explore recent advances in image analysis using deep neural networks for object detection, segmentation and classification.
Among the functions that compose the role Cartogalicia will play within the project POSEIDON, are the following:
- Design and development of a multi-sensor processing and fusion software based on an open architecture that allows the system to abstract the specific hardware imposed by the selected UAV platform, thus facilitating the reconfiguration of the system in function of SWAP requirements (Size, Weight And Power). This system will integrate the two sensoring technology proposed in the project: Polarimetric radar and EO/IR optical sensors.
- Design and development of an EO/IR image Analysis system for the detection and classification of waste discharges and marine pollution. In particular, the combination of a thermal sensor and a multispectral sensor in the VIS/NIR range is proposed. It also proposes the application of the recent advances in deep neural networks for the intelligent analysis of the images in order to detect objects in the sea and the segmentation and classification of landfills.
- Study of data fusion techniques for a joint exploitation of polarimetric and EO/IR radar data and development of a multisensor Fusion software module based on these techniques.
- Experimental validation of the system, through its integration aboard aerial means, and the carrying out of tests that allow to demonstrate the efficiency of the system, as well as the viability of its operation from small UAVs.
Link of project: http://poseidon.cnit.it/
Este proyecto está financiado por el CDTI y cofinanciado por el programa de investigación e innovación Horizonte 2020 de la Unión Europea en el marco de ERA-NET Cofund MarTERA (Tecnologías marítimas y marinas para una nueva era).