In the analysis, object-oriented approach is preferred as an image processing method since it combines the spectral information with spatial information as shape, size,texture andneighborhood relations to increase the classification abilities.
The methodical principles of object-oriented image processing consist of two basic domains: the segmentation and the classification i) Segmentation which aims at producingbasic processing units for object oriented image analysis (so called image objects orsegments). The resulting image objects are the image units instead of pixels and have attributes as not only spectral statistics but also shape information, relations toneighbouring objects and texture ii) Classification which is based on the class-hierarchy. The classhierarchy contains the classification rules to which the image will be classified. Class description is defined and a network of hierarchy is structured by relations of sub and super objects through levels for each class. In this study, the Definiens Professional Earth software (Ver. 5.0) is used in the analysis.