On-line system for monitoring and forecasting Earth surface changes using sequences of remotely-sensed imagery

Irregular temporal sampling is a common feature of geophysical and biological time series in remote sensing. This study develops an on-line system for monitoring and forecasting ground surface changes by adaptively generating an appropriate synthetic time series at regular interval with recovering missing measurements for sequential images that are compiled at irregular time interval from Earth ground. The proposed system integrates an adaptive reconstruction technique and a classification algorithm. The reconstruction method incorporates temporal variation according to physical properties of targets and anisotropic spatial optical properties into image processing techniques. The classification algorithm segments each image frame by partitioning into physically meaningful regions whose statistics are expected to depend on the physical properties of the region. This adaptive approach allows successive refinement of the structure of objects that are barely detectable in the observed series and monitoring of temporal variation in surface characteristics by observing statistical changes between contiguous image frames in the adaptive system