ACOSOP - Automatic COntrol System of Offshore Platforms through satellite remote sensing and development of a GIS-based application for coastal marine environmental monitoring

Over the years, offshore hydrocarbon extraction worldwide has grown constantly, linked to an incessant increase in energy consumption, connected to a process of urban and industrial development that has shown no signs of slowing down over time. This prospect of economic growth has encouraged oil companies to search for new hydrocarbon deposits in the oceans and at ever greater depths, where there are still largely unexplored areas. Offshore oil and gas drilling and extraction activities require the construction of installations to support the offshore oil extraction industry, and since the 1950s until today, thousands of platforms have been installed around the world. Most of these installations are located in coastal marine environments, within highly vulnerable and therefore fragile ecosystems, which can suffer significant damage in terms of ecological wealth and economic resources in the event of accidents. It is therefore important that the environment is preserved by taking appropriate measures for its protection and response strategies, even in the event of major accidents. Italy, with its long history in the oil and gas industry, has always been interested in the exploitation of offshore fields, and Italian offshore platforms, which are the subject of this PhD thesis, are mainly distributed in the northern and central Adriatic Sea (there are 121 of them) and are located near the coast, at depths of between 10 and 120 metres. The presence of offshore structures installed in shallow waters a few kilometres from the Italian coast requires continuous surveillance, which is very resource-intensive. Some of these structures were built over thirty years ago and are at risk of deterioration, but in the event of an accident, they all represent potential sources of pollution with a significant environmental impact on the marine ecosystem. Climate change and the increase in extreme weather events associated with it also increase the vulnerability of these infrastructures. Such events increase the environmental risk as they can lead to potential accidents, such as the one that occurred a few years ago when a Croatian platform was completely dislodged by a severe storm. The use of remote sensors, such as those carried by Earth Observation (EO) satellites, is a particularly effective and economical monitoring tool, as it allows for the control of the status of platforms with a synoptic and repeated view over time, especially when a large number of satellite platforms are used for monitoring. These considerations demonstrate how important it can be to systematise different tools for monitoring offshore installations based on Earth observation data. This objective can be pursued through the design and development of an application that automatically controls offshore platforms and allows the environmental conditions of the surrounding marine areas to be verified, representing an important tool for identifying potential disasters and supporting decisions during operational management by choosing the most appropriate responses in order to reduce short- and long-term environmental impacts and socio-economic losses. It is essential, especially during emergency management phases, to have constantly updated and timely information in order to protect the environment, the economy and the safety of the local population. Through various satellite data processing algorithms, it is possible to collect information on oil facilities and the environmental characteristics of their surroundings, and to update this information with the observation frequency offered by the satellites currently in operation used for monitoring. The research project, which will be presented in this thesis, was developed in collaboration with the Italian Space Agency (Agenzia Spaziale Italiana - ASI) as part of the National PhD in Earth Observation at Sapienza University of Rome. Its main purpose was to develop and test an integrated control system designed as a GIS (Geographic Information System) application and implemented through a workflow consisting of automated routines which, starting from the acquisition and processing of Copernicus Sentinel-1 satellite data, identify platforms, verifies any absences and allows the automatic sharing of analysis results both through cartographic products sent by e-mail and through a graphical interface accessible via the web on various devices. The Sentinel-1 satellite was chosen because, operating continuously over the Mediterranean, it offers constant monitoring frequency of 7 days in ascending mode and the same in descending mode. Furthermore, being a C-band SAR satellite, it is not hampered by cloud cover or night-time hours. As a specific objective in this research project, a tool for the automatic detection of offshore platforms (object detection) has been developed. Through the systematic analysis of Sentinel-1 SAR satellite images using deep learning techniques integrated into ArcGIS Pro, this tool can verify the presence of offshore platforms and alert users in the event of their absence with a warning notification. The design of a WebGIS application on the ESRI platform has enabled the creation of a demonstrative downstream service, which allows the communication and sharing of geo-information with potential operators. Once developments are complete, a comprehensive suite of monitoring tools will be available, which will enable more informed decision-making regarding the measures to be taken in response to emergencies and, where possible based on the operational characteristics of Earth Observation (EO) missions, synergistic near real-time monitoring with dynamic visualisation of relevant environmental parameters.