Multi-temporal classification for irrigation detection in the Vinalopo region in Spain using ASTER images

There is a growing interest in monitoring land usage using remote sensing data. Besides, sometimes it is very important to find out irrigation patterns in order to have additional data for local water management agencies. That applies in the basin of river Vinalopo in Spain, where water is traditionally a limited resource. However, this area presents special severe difficulties when satellite images are exploited. The typical small parcel size and the particular irrigation practice that only applies in the drought months present a challenge in a classification process. In this paper, some multi-temporal classification techniques are used to classify crops in the area of Novelda using ASTER images. Several pre-classification techniques and multi-temporal classification methods are compared in this study. The results show that some techniques provide promising results for detecting irrigated crops taking into account the adverse characteristics of the area.