• clustering algorithm
• real time application
• spatial application
• lightning imager

This thesis deals with the implementation of a new kind of clustering algorithm that will be used by a Lightning Imager (LI) mounted on ESA meteorological satellites of third generation (MTG). The development of this algorithm is commissioned to the Department of Information Engineering by Selex ES in cooperation with ESA (European Space Agency). At first, the algorithm is designed to meet the stringent timing requirements. Then, the new clustering algorithm is implemented on MATlab and VHDL languages.
The thesis is organized as follows:
• Chapter 1, Meteorological satellites, introduces the principal orbits used by every kind of satellite, the data provided by meteorological missions and the history - past, present and future - of meteorology. A particular section is reserved to EUMETSAT and its missions with a large presentation of EUMETSAT three generations of satellites:.
o Meteosat First Generation series. A short resume of history, position, lifetime and technical data with a particular section dedicated to the principal instrument on board, the MVIRI;
o Meteosat Second Generation, the active series in orbit around the Earth. It gives details about satellite principal instruments, SEVIRI and GERB, and their characteristics;
o Meteosat Third Generation. A presentation about the new MTG series that will be soon in orbit. It explains which instrument will be on board and which characteristics they will have. Also, the difference between MTG-I and MTG-s will shown;
• Chapter 2, Lightning Imager, exposes what is an imager and the requirements necessary to implement this device. Then, it presents the problems during the capturing and which methods are used to increase the quality of events detection;
• Chapter 3 is focused on the LI technology and his most important parts: requirements will be analyzed and their feasibility will be evaluated. The next step is the design of the clustering algorithm and the work will be concluded with three output codes: a MATlab High-Level (MHL) model, a MATlab Bit-True (MBT) model and a prototype of VHDL model for hardware implementation.