English 
Italiano 
In metallic devices, various forms of radiation collapse are one of the major causes of plasma degradation leading to disruptions. Some of the most advanced scenarios, with impurity seeding or detachment, are prone to developing excessive edge cooling, which very often ends with the collapse of the discharge. This type of disruption is also typically preceded by the formation of an instability called Multifaceted Asymmetric Radiation From the Edge (MARFE). This manifests itself as a blob of radiation above the X point or in various other regions of the plasma edge. This work reports on the development of image processing algorithms to detect MARFE as early as possible, to contribute to disruption prediction. Particular attention has therefore been devoted to general techniques, which can identify the features typical of MARFEs in any region of the plasma cross-section. The developed tools are sufficiently flexible to allow selecting a good trade-off between sensitivity and specificity. The solutions adopted are also all compatible with the implementation in real time networks. The examples provided have been collected using JET cameras during the H and T campaigns, run in preparation of the recently concluded JET campaign in D-T. Given the warning times achieved, it is expected that the early and accurate detection of MARFEs can contribute significantly to the prevention of the disruptions caused or preceded by these instabilities.
In a world increasingly facing new challenges at the forefront of plasma scientific research and technological innovation, CNR and ISTP pledge progress and achieve an impact in the integration of research into societal practices and policy
