The project consists of the four following parts: Part 1 Automatic image analysis and machine learning Main Aim: For proper analysis and use of pictures, there is a need for a system to help humans to pinpoint/select issues of interest to be further inv estigated. This could be 1) specific objects of interest, e.g. coverage of corals or leaks of hydrocarbons or 2) changes in patterns of vast amount of data on one geographical location e.g. gradually present or absent of specific species. Part 2 Combini ng multi sensor data for optimal benefit of information Main Aim: is to use one or several of the identified case studies in the IEM project and the AUR-LAB cruises to identify which key-environmental variables and parameters that should be measured in g iven situations (1-3 described below), and use multivariate analysis to interpret the results with respect to the source for observed change and the co-variance between variables and parameters. Part 3 How to use processed data into a decision making pr ocesses? Main Aim: The focus in this part will be on finding the optimal way of presenting the data types, or a set of these, described in part 1 and 2. Furthermore, as basis for environmental management the aim is to identify what kind of data different decision makers needs and if some of the data need to be shared between the different decision makers. Part 4 Piloting process Main Aim: The focus in this part is to apply Part 1-3 on an actual offshore case. This part will also be a validation of the s olutions chosen in part 1-3. The pilot will be specified by the IEM project. I addition to reflecting key issues identified as success factors in the IEM project, it is also believed that the different parts will give valid knowledge and experiences on how to optimise and use environmental data in management in general.
Project leader: Vidar Hepsø
Institution: EQUINOR ASA