Object-Based Image Analysis

Object-based image analysis (OBIA) is a new approach to the analysis of remote-sensing (RS) images in which the basic units, instead of being individual pixels, are image-objects. An image-object is a relatively uniform region of a digital image somewhat different from its surroundings. Typically, image-objects are initially delineated by a segmentation algorithm and then further aggregated using a semiautomated classification that takes into account both their internal features and their mutual relationships. OBIA aims to replicate visual interpretation through semiau-tomated means, providing greater integration with vector geographic information systems (GIS) with increased consistency and efficiency.

OBIA is an increasingly popular alternative to the traditional pixel-based approach to semiau-tomated thematic mapping using RS imagery. Pixel-based methods assume that different thematic classes behave like distinct surface materials susceptible ...