Crop type mapping in a highly fragmented and heterogeneous agricultural landscape: A case of central Iran using multi-temporal Landsat 8 imagery

Crop type mapping and studying the dynamics of agricultural fields in arid and semi-arid environments are of high importance since these ecosystems have witnessed an unprecedented rate of area decline during the last decades. Crop type mapping using medium spatial resolution imagery data has been considered as one of the most important management tools. Remotely sensed data provide reliable, cost and time effective information for monitoring, analyzing and mapping of agricultural land areas. This research was conducted to explore the utility of Landsat 8 imagery data for crop type mapping in a highly fragmented and heterogeneous agricultural landscape in Najaf-Abad Hydrological Unit, Iran. Based on the phenological information from long-term field surveys, five Landsat 8 image scenes (from March to October) were processed to classify the main crop types. In this regard, wheat, barley, alfalfa, and fruit trees have been classified applying inventive decision tree algorithms and Support Vector Machine was used to categorize rice, potato, vegetables, and greenhouse vegetable crops. Accuracy assessment was then undertaken based on spring and summer crop maps (two confusion matrices) that resulted in Kappa coefficients of 0.89. The employed images and classification methods could form a basis for better crop type mapping in central Iran that is undergoing severe drought condition. (C) 2016 Elsevier B.V. All rights reserved.