ABSTRACT
Object oriented classification, image extraction, or Object Based Image Analysis (OBIA), represents a more automated method of image classification than the previously examined traditional classification approaches (unsupervised classification and supervised classification). In an OBIA approach, in addition to spectral properties, the software uses image-based “cues” that are similar to a human’s process for visual interpretation (such as color/tone, texture, size, shape, shadow, site/situation, pattern, and association). This exercise will explore the use of an OBIA approach to select user-defined training samples, or objects, for creating an object-oriented image classification to classify land use and land cover.
