Unsupervised Sequential Classification of Modis Time-Series

In this paper we present a novel hypertemporal unsupervised sequential classification algorithm. We illustrate the usefulness of this algorithm at the hand of a case study. For our case study, we consider a Moderate Resolution Imaging Spectroradiometer dataset containing two prominent South Afrcian land cover classes, namely natural vegetation and settlement. For the case study we considered in this paper, the proposed sequential classification approach performed on average 1.33 times better than conventional k-means.