The study is conducted using historical ground truth data of WND cases derived by several epidemics that have been affecting the Italian territory since 2008. WND outbreaks, with cases reported in mosquitoes, birds and horses are collected from the National Information System for Animal Disease Notification (SIMAN). EO data derive from different sources (Sentinel-2, Sentinel-3, PROBA-V, etc.), pre-processed and harmonised.
Definition of requirements
Information regarding EO data to be used, criteria to select ground truth data, temporal interval to be analysed and different Deep Neural Network models are evaluated and defined. Selection criteria and preparation of remotely sensed products are then investigated, considering data from multiple sources, various sensors, spectral bands, spatial resolutions and revisit times. WND and EO data are selected to guarantee a correct spatial and temporal representation of the last ten-year epidemics.
Data retrieval and processing
WND cases are extracted from the official repository of the Italian Ministry of Health (National Information System of Animal Disease Notification - SIMAN), validated and selected, in space and time, according to the requirements defined in the previous phase.
WND ground truth outbreaks are split in different datasets to be used to train and test the DNN model, then fine-tune the model and hence make predictions and evaluate the overall accuracy. Selected EO data are collected from different sources and stored in a centralised system that organises and pre-process them according to the requirements. Classical statistical models for WND spread (suitability analysis, logistic regression, etc.) are developed to be compared with AI model performance.