Detecting fog using machine learning on satellite data has been researched before, but not on a global scale using syn thetic data. The aim of the thesis is to use a synthetic dataset of simulated MODIS satellite data to determine the viability of machine learning algorithms for detecting fog in satellite images. The synthetic dataset we use is simulated using a fast ra diative transfer model called RTTOV by inputting various atmospheric information for different conditions. The dataset is tabular and no spatial or temporal relation ship exists between the data points meaning each pixel is treated independently. We use the synthetic data to train and evaluate numerous machine learning models including various implementations of XGBoost and feed forward deep neural net works. We demonstrate that classification models can achieve good recall values on synthetic data when oversampling fog in the training data, the best being 0. However, we find that this comes at the cost of a large amount of false positives evident by the low precision value of 0.
Aktuella examensarbeten Nedan finner du en lista över möjliga områden inom vilka examensarbeten kan definieras. De anknyter samtliga till aktuell forskning och kan anpassas till att innehålla en större del av teoretiskt arbete inom modell- och algoritmutveckling eller en större del av numeriskt orienterad algoritmutveckling och utvärdering. Together we have generated important knowledge about maintenance processes regarding aircraft engines, in particular the jet engine RM We have especially constructed mathematical models that describe the problem to find an optimal maintenance plan for a whole engine during a contract period, taking into account the existence of used parts with shorter lives at warehouses, work costs, and possible side constraints on the engine condition at the end of the contract period. In a master thesis project, ideally for two collaborating students, we wish to investigate a number of possible solution methodologies for the maintenance problem. It is a very large scale linear mixed integer optimization problem, for which several possible lines of attack are potentially interesting. In a collaboration with us you will implement and evaluate them on realistic data provided by VAC.