Abstract
Urbanized areas are characterized by high complexity and diversity caused by repeating of many features – natural and man made. Object Based Image Analysis enables to reduce image complexity by its distribution into several levels of homogenous segments.
It also operates with many tools for effective feature classification based not only on pixel values but also on feature shape, texture and etc. Object based classification of VHR QuickBird satellite imagery enabled to develop an effective classification method based on fuzzy model with membership functions for urbanized areas.
Applicability of the model for different areas of the same QuickBird image has been proved. Though, the model is not applicable for suburban areas with high share of commercial suburbanization land use/cover categories without significant improvements.
Further improvements like elevation value classification rule implementation for building features have been recommended.