Ongoing projects

Coastal SAGE

Coastal erosion is an unrelenting phenomenon which is of importance to the Maltese Islands as the coast is one of the most-intensely used and visited areas. Research in the downstream Earth Observation sector is key to achieving reliable and cost-effective monitoring of coastal erosion. Persistent Scatterer Interferometry (PSI) techniques utilize Synthetic Aperture Radar (SAR) onboard satellites to provide millimetric deformation estimates. However, SAR suffers from speckle noise, which can affect the PSI processing pipeline and the resulting deformation maps and their interpretation. The “Coastal Satellite Assisted Governance (tools, techniques, models) for Erosion” (Coastal SAGE) project will use image processing and deep learning techniques to address two key aspects of the PSI pipeline: denoising of interferometric phase and phase unwrapping. The developed denoising and unwrapping methods will be used to extract ground displacements from time series of synthetic aperture radar (SAR) acquisitions, in order to estimate deformation and displacement in study areas around Malta and Gozo. These estimates will be validated through in-situ sensors. The Coastal SAGE project is led by Dr Ing. Gianluca Valentino from the University of Malta, with the Marine and Storm Water Unit of the Public Works Department within the Ministry of Transport, Infrastructure and Capital Projects as a partner within the consortium, led by Dr George Buhagiar.

Duration:  2020-2022 

Grant: €150,000  

WAter Resource Management platform using Earth Observation (WARM-EO)

Malta has a semi-arid Mediterranean climate with a low availability of natural renewable water resources for sustaining the production of drinking water, the water demand of agriculture and the sustainability of the environment. This coupled to a high population density exerts a significant pressure on natural water resources and associated ecosystems. The overall objective of the WARM-EO project is to develop a Water Resource Management platform that can be used to estimate irrigation water consumption of particular crops at country level. Nevertheless, models adopted in other countries are not applicable to small Mediterranean countries such as Malta where the parcels are too small and fragmented for the resolution of existing open-access satellite services. The WARM-EO project aims to develop a cost-effective high-resolution evapotranspiration model to estimate the irrigation water use at parcel level. More specifically, the project aims to develop a deep-learning based multi-frame superresolution algorithm to improve the resolution of Sentinel-2 optical images to enable the computation of vegetation indices at 3 m resolution. Moreover, this project intends to fuse in-situ data obtained from nine farms scattered around the island and remote sensing data to estimate land surface temperature at parcel level. The resulting high-resolution evapotranspiration model will be validated against ground-truth evapotranspiration measurements provided by the Energy and Water Agency as part of the CF.PA 10.0096 Project. A Web-GIS service will also be developed to estimate the evapotranspiration of the whole country.

Duration:  2020-2022 

Grant: €150,000 

VideO Light field Acquisition and REstoration (VoLARe)

The state-of-the-art in augmented-reality (AR), virtual-reality (VR) and cinematic recording is the Lytro Immerge 2.0, using a rig of ninety-five (95) cameras. This technology allows digital refocusing after capture and facilitates 3D modelling and the realistic integration of computer-generated content. The major hurdle is that this system is too expensive for most productions. The VoLARe project involves the design and development of a low-cost video light field capturing prototype which aims to reduce the number of cameras by a factor of nine (9) thus reducing the throughput down to 30 Gbps. The major challenge in this project is to develop spatial, angular and temporal restoration techniques to improve the quality of the captured video light field. Our partners are Stargate Studios Malta, a video production and visual effects company that will use the developed video light field capturing system in a production. 

Duration:  2020-2022 

Grant: €200,000 

Visit the website  

SAtellite TraIning and NETworking (SATINET)

The Pléiades satellite is the first satellite that captures three consecutive high resolution images of the same region within a relatively short time-frame. This enables researchers to use computer vision techniques to estimate the elevation of objects from the land surface. Researchers at CNES have developed a processing pipeline that is able to generate a 3D reconstruction of scenes captured by the Pléiades satellite. This technology makes it possible to update the digital elevation models on a daily basis. Nevertheless, the 3D reconstruction method has an inherent error margin and can also introduce distortions in the models. The aim of the SATINET project is to use images captured by Synthetic Aperture Radar from Sentinel-1 and adopt more advanced computer vision techniques to improve the quality of the existing 3D reconstruction model. This project will also finance an M.Sc. by research project in collaboration with CNES within the Department of Communications and Computer Engineering at the University of Malta to work on Crop classification. 

Duration:  2019-2020 

Grant: €50,000 

Face Image Restoration using Deep learning (Deep-FIR)

The Deep-FIR project aims to design and implement a face image restoration algorithm that is able to restore very low-resolution facial images captured by CCTV systems with unconstrained pose and orientation. Apart from improving the quality of the restored facial images, this project intends to reduce the complexity, and therefore the time needed to enhance an image or video frame. The developed algorithm will be tested on real-world CCTV videos and compared against existing video forensic tools used by forensic experts in their labs. Project Deep-FIR financed by the Malta Council of Science & Technology, for and on behalf of the Foundation of Science and Technology, through the FUSION: R&I Technology Development Programme.


Duration:
 2018-2021

Grant: €200,000

Visit the website

The ALICE Experiment @ CERN

The ALICE experiment at the Large Hadron Collider at CERN studies the quark gluon plasma thought to have existed at the start of the Big Bang. Due to the high interaction rate in the experiment resulting from the collisions of heavy-ion particles, particle tracks produced in ALICE's detectors are often characterised by high occupancy and high noise. Members of the Data Science Research Group (DSRG) collaborate with two such detectors. In the case of the Time Projection Chamber, the objective is to identify low-momentum (low radius) helices which are not useful for physics analyses. On the other hand, in the High Momentum Particle IDentification detector, the aim is to identify elliptical patters from which the Cherenkov angle can be calculated.

 

 

 

 

 


https://www.um.edu.mt/platform/dsrp/ongoingprojects