engleski

Object-based LULC classification of urban and peri-urban areas

During the last decade significant attention has been paid to object-based image analysis (or object-oriented image analysis) methods as the means for the land cover and land use classification of data acquired by remote sensing. This paper will present object-based LULC classification performed on multispectral remote sensing imagery of urban and peri-urban areas. Emphasis in the classification process will be placed on the mapping of the impervious surfaces as opposed to vegetation, soil, and water surfaces because that land cover class is made out of the most noticeable features of urban spaces: buildings and transportation network. Impervious surfaces, or more specifically their quantity and spatial pattern, are the ones that define the human settlements morphology and structure most clearly and have undeniable influence on the water and energy flux in the environment. Remote sensing data used for this research will be RapidEye satellite imagery that comes in 5 spectral bands (red, green, blue, Red Edge, NIR) with the spatial resolution of 5 m and with the 16-bit radiometric resolution. Error matrix will be produced by comparison of the classification results and very high spatial resolution remote sensing imagery (available through services such as GoogleEarth) that will serve as the ground truth data. Software used for the object based image analysis and most of the other tasks will be Trimble eCognition Developer 8 and ESRI ArcGIS Desktop 9.3.1 (ArcInfo).

object-based image analysis; LULC classification; multispectral remote sensing data; urban areas; impervious surfaces

nije evidentirano