Object-based image analysis

Object-based image analysis for remote sensing data is built on two main steps: (1) image segmentation, which aims to partition an image into homogeneous objects and (2) classification, which assigns a thematic class to the partitioned objects. This workflow is one of the main pillars of my research, aiming for automated methods to turn data into information with multiple applications, like agricultural mapping from time series data, land use-land cover mapping, drone imagery interpretation or geomorphology.