GEO-K takes part to the study for super-classification of multispectral data

GEO-K participates to the study for the super-classification of Landsat Multispectral products. The study will be published on the IEEE Journal of Selected Topics in Applied Earth Observation and Remote Sensing in the next period.

What is super-classification?

With it, we consider the ability to discriminate the different “materials” that form a single pixel signal. Often these materials are different and composite on the Earth’s surface. With this approach, we can ultimately increase the capability of satellite instruments in terms of spatial resolution.

The methodology proposed by GEO-K and the Earth Observation Laboratory of the University of Rome Tor Vergata is based on advanced techniques exploiting Neural Networks.

The first results obtained by the team were presented at the URBAN 2015 Remote Sensing Joint Event in Losanna, Switzerland.