Open boundary dynamic time warping for satellite image time series classification

In this paper, we present weighted version of the Dynamic Time Warping algorithm for land cover and land use classification. The method is flexible to account for multiyear crops and double cropping, tropical forests, and pasture using few training samples. In our case study the algorithm had global accuracy of 87% showing results in line with other land use/cover products. We consider the method suitable to make land cover and land use maps using data sets such as the EVI time series from the MODIS sensor.