Section: New Results

Discovering Evolution Patterns from Satellite Image Time Series

Participant : Florent Masseglia.

Satellite Image Time Series (SITS) provide us with precious information on land cover evolution. By studying these series of images we can understand the changes of specific areas but also discover global phenomena that spread over larger areas. However, discovering these evolution patterns implies to consider two main challenges, related to the characteristics of SITS and the domain's constraints. First, satellite images associate multiple measures with a single pixel (the radiometric levels of different wavelengths corresponding to infra-red, red, etc.). Second, these evolution patterns spread over very long periods and they may have different start time and end time depending on the region. Furthermore, the non evolving regions, which are majority and dominate over the evolving ones, challenge the discovery of these patterns. In [35] , [47] we propose a SITS mining framework that allows for discovering these patterns despite these constraints and characteristics. Our proposal is inspired from sequential pattern mining and provides a relevant visualisation principle. This work has been accepted to the EGC 2011 conference (january 2011).